TMD DISCUSSION PAPER NO. 19 RICE PRICE POLICIES IN INDONESIA: A COMPUTABLE GENERAL EQUILIBRIUM (CGE) ANALYSIS Sherman Robinson Moataz El-Said International Food Policy Research Institute Nu Nu San Winrock International International Food Policy Research Institute With Achmad Suryana Hermanto Dewa Swastika Sjaiful Bahri Center for Agro-Socioeconomic Research (CASER) Trade and Macroeconomics Division International Food Policy Research Institute 2033 K Street, N.W. Washington, D.C. 20006, U.S.A. June 1997 TMD Discussion Papers contain preliminary material and research results, and are circulated prior to a full peer review in order to stimulate discussion and critical comment. It is expected that most Discussion Papers will eventually be published in some other form, and that their content may also be revised. This paper was prepared for the IFPRI-CASER project “Modeling the Future of Indonesian Agriculture.” Abstract This paper presents an agriculture sector focused Computable General Equilibrium (CGE) model for analyzing the economy-wide impacts of changes in production technology, protection, and market structure on resource allocation, production, and trade in Indonesia. The paper incorporates a specification of the rice market and models Bulog's (National Logistic Agency) behavior using a mixed complementarity approach. This approach allows the specification of inequalities and changes in policy regime as prices and/or stocks move within specified bands. The model is used to examine the impact on the Indonesian economy of changes in rice yields given different assumptions about the operations of Bulog. The general equilibrium approach does capture and quantify the effects of the price support policies on resource allocation, trade, relative prices, and the government budget. An important result is the inefficient allocation of resources within the agriculture sector and the rest of the economy if Bulog operates to maintain the rice price when there are significant increases in rice productivity. Instead of releasing resources to other high-value agriculture uses and non- agriculture uses, the price support scheme attracts more resources into rice production. In addition, the price support program is costly and strains the government accounts, even if the administrative cost of operating the program are ignored. ii Table of Contents 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 2. Price Policies and Operational Structure of Bulog . . . . . . . . . . . . . . . . . . . . . . . . . . 3 2.1 Rice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 2.2 Soybean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 2.3 Sugar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2.4 Wheat and wheat flour . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 2.5 Funding and expenditure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 3. The Social Accounting Matrix (SAM) for Indonesia . . . . . . . . . . . . . . . . . . . . . . . 12 4. Equations of the Core CGE Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 4.1 Price Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 4.2 Quantity Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 4.3 Income Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 4.4 Expenditure Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 4.5 Market Clearing Conditions and Macroeconomic Closure . . . . . . . . . . . . . 26 5. Base Solution, Policy Experiments, and Results . . . . . . . . . . . . . . . . . . . . . . . . . . 31 5.1 Structure of the Economy: Base Solution . . . . . . . . . . . . . . . . . . . . . . . . . . 31 5.2 Policy Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 5.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Rice Productivity Decline: Experiment 1 . . . . . . . . . . . . . . . . . . . . . . . . 34 Rice Productivity Improvement: Experiment 2 . . . . . . . . . . . . . . . . . . . . 37 Rice Productivity Improvement Without Bulog Intervention: Experiment 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 6. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Appendix 1 Supplementary Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Appendix 2 The AG-CGE Model: GAMS code . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 Appendix 3 The Disaggregated SAM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 iii List of Tables Table 3.1. An Aggregate Social Accounting Matrix (SAM) for Indonesia, 1990... 15 Table 3.2. SAM disaggregation (activities, commodities, factors, and institutions) 16 Table 4.1 Definition of model indices, parameters, and variables.......................... 18 Table 4.2 Price equations........................................................................................ 20 Table 4.3 Quantity equations.................................................................................. 23 Table 4.4 Income equations.................................................................................... 25 Table 4.5 Expenditure equations............................................................................. 28 Table 4.6 Market clearing and macro economic closures....................................... 29 Table 5.1 Structure of the Indonesian economy, 1990, the base year for the model 34 Table 5.2 Government accounts, rice productivity decline.................................... 37 Table 5.3 Rice prices and quantities, rice productivity decline.............................. 38 Table 5.4 Macro results, rice productivity decline.................................................. 38 Table 5.5 Government accounts, rice productivity improvement........................... 40 Table 5.6 Rice prices and quantities, rice productivity improvement..................... .41 Table 5.7 Macro results, rice productivity improvement........................................ 41 Table 5.8 GDP deflators with and without Bulog intervention with a 25% improvement in rice productivity........................................................... 45 Table 5.9 Changes in real and nominal value added shares with a 25% rice productivity improvement....................................................................... 45 Table A.1.1 Production and quantity in Bulog market operations for paddy and rice, 1969 - 1995.............................................................................................. 52 Table A.1.2 Production and quantity in Bulog market operations for sugar, 1970 - 1994........................................................................................................ 53 iv Table A.1.3 Production and quantity in Bulog market operations for soybean, 1970 - 1994........................................................................................................ 54 Table A.1.4 Paddy and rice prices, 1969 - 1995......................................................... 55 Table A.1.5 Soybean prices, 1977 - 1993.................................................................. 56 Table A.1.6 Sugarcane prices, 1970 - 1993............................................................... 57 Table A.2.1 Definition of model indices, parameters, and variables......................... 60 Table A.3.1 A descriptive SAM for Indonesia........................................................... 77 Table A.3.2 Micro SAM for Indonesia: 1990............................................................ 78 List of figures and charts Chart 1 Bulog market operations for paddy and rice.......................................... 4 Figure 2.1 Government support floor price and average producer price for paddy, 5 1974 - 1993.............................................................................................. Chart 2 Bulog market operations for soybean..................................................... 8 Chart 3 Bulog market operations for white sugar............................................... 9 Chart 4 Bulog market intervention for wheat and wheat flour............................ 11 Figure 5.1 Changes in the value of non-agriculture and agriculture production with rice productivity improvement............................................................... 43 Figure 5.2 Changes in the value of non-agriculture and agriculture imports with rice productivity improvement...................................................................... 43 Figure 5.3 Changes in the value of non-agriculture and agriculture exports with rice productivity improvement...................................................................... 43 Figure 5.4 Changes in the value of rice, fruit and vegetables, and other agriculture production with rice productivity improvement..................................... 44 v List of Abbreviations AG-CGE Agriculture Sector Focused Computable General Equilibrium BPS Biro Pusat Statistik Bulog National Logistic Agency CES Constant Elasticity of Substitution CET Constant Elasticity of Transformation CGE Computable General Equilibrium COL Jakarta Cost of Living Index Dolog Regional Logistic Agency GAMS General Algebraic Modeling System KUDs Village Cooperatives SAM Social Accounting Matrix For Example, when there was a sudden chili price hike in February 1996, Bulog imported chili from1 Thailand and sold it in local markets. 1 1. Introduction Food policy in Indonesia aims to achieve food security by increasing food production, raising farm income, improving nutritional status of the people, and to ensure the availability of food supplies at affordable prices (Bulog, 1995). For the last 27 years, Indonesian food policy has centered on rice, the most important staple crop and sometimes referred to as a strategic crop. Maize, soybean, sugar cane, and cassava are considered important secondary crops. Since the early 1970s, rice policy in Indonesia have sought to attain food self sufficiency through price support policies, price stabilization policies, and public investment policies (Pearson et al., 1991). Bulog (national logistic agency) was authorized to implement the pricing policies for rice and to provide monthly rations to the military and civil service. Bulog also provides assistance in releasing food from stocks in case of national catastrophes such as earthquakes and floods. Bulog’s market interventions were later extended to a wide variety of commodities including maize, mungbean, sugarcane, soybean, soybean meal, wheat, wheat flour, chicken, and eggs. Currently, rice, sugarcane, garlic, soybean, wheat, wheat flour, and crude palm oil are included on the list. In addition, Bulog occasionally operates in other commodity markets, especially when the price fluctuates extensively due to shortages or market imperfections .1 Bulog's intervention to achieve commodity price stabilization has been acclaimed for its contribution to Indonesia's political stability and development. With an average annual growth rate of 6.7 percent, one of the fastest growing economies in Asia, Indonesia's dependence on agriculture has declined. The contribution of agriculture to GDP is nearly 23% in 1995 compared to 48% in the early 1970s. In addition, as the international economy moves toward trade liberalization, multilateral trade agreements have emphasized reduction of government protection in the agriculture sector. Consequently, the debate in Indonesia on market interventions has been accelerated in recent times. In order to assess the economy-wide impacts of commodity market interventions, this study presents an Agriculture sector focused Computable General Equilibrium (AG-CGE) model for Indonesia. This analytical framework focuses on agriculture and on links between the agriculture and non-agriculture sectors. The model provides a good framework for analyzing the impacts of changes in production technology, protection, subsidies, and market structure on resource allocation, production, employment, and trade. The model used in this paper incorporates a specification of the rice market and the role of Bulog, and is used to examine how changes in rice yields affect the economy under different scenarios concerning Bulog’s management of the market and trade. 2 In the next section, we discuss the operational structure of Bulog for long-term protected commodities: rice, soybean, sugar cane, and wheat. Section 3 introduces the Social Accounting Matrix (SAM), which incorporates much of the data used in the model. Section 4 outlines the equations of the core CGE model and a specification for Bulog operations in the rice market. Section 5 discusses model calibration, policy experiments, and their results. Materials in this section are largely drawn from Bulog: The National Grain Authority of Indonesia2 (1992) Jakarta Indonesia, and Bulog: National Logistic Agency (1995) Jakarta. Indonesia. These include the six most prevalent rice varieties found in the latest bi-monthly market survey by3 the Census Bureau. 3 2. Price Policies and Operational Structure of Bulog2 2.1 Rice The basic concepts underlining the current price level policies for rice were laid out by Mears and Afiff in 1969. The four major policy objectives are: (1) setting the floor price high enough to stimulate production, (2) establishing a ceiling price which assures a reasonable price for consumers, (3) maintaining sufficient range between these two prices to provide traders and millers reasonable profit after holding rice between crop seasons, and (4) keeping an appropriate price relationship between domestic and international markets. Bulog’s implementation of these price and price stabilization policies for rice involves setting a floor price and a ceiling price, procuring paddy or milled rice, stock management and quality control, distribution, as well as importing and exporting. The floor price and ceiling price are determined by Bulog in conjunction with three ministries: Coordinating Ministry of Economics, Ministry of Trade, and Ministry of Agriculture. During the last decade, floor price decisions were based on simulation analyses drawing on estimates of elasticities and other parameters. The impact of different combinations of floor prices and fertilizer prices on farmers’ income, inflation, consumer prices, and farmers’ term of trade were analyzed and “appropriate” price levels were then selected (Amang, 1993). The floor price for paddy is usually announced during October to December for the following calendar year. However, the ceiling price has not been officially announced since 1980, yet experienced private traders said to have been able to project the price that Bulog defends. Before 1979, the Jakarta cost of living index (COL) was used as the primary criterion for setting the ceiling price. During the year, Bulog would occasionally release stocks to limit the price increases of the bundle of COL rice varieties so that the changes stay within the rate of inflation. In April 1979, when consumer price indices (CPI)3 for 17 major cities were introduced, the CPI become the new basis for setting the ceiling prices (Mears, 1981). The margin between the floor price and consumer retail price fluctuates over time (Table A.1.4.). The average margin in the 1990s has been approximately 21%. Chart 1 illustrates Bulog’s market operations for paddy and rice. Village cooperatives (KUDs) were established in 1973 with one of their functions being the purchase of paddy from farmers. Dolog (regional logistic agency) pays the floor price plus a commission for the KUDs services in purchasing paddy from farmers. If KUDs are pressed beyond their 4 Farmers Millers and tradersDOLOG task forceCoooperative (KUDS)Farm floor price DOLOG (BULOG)Imports KUD floor price Coop & private authorized distributors Government agencies DOLOG task force Private wholesalers Rerail markets Military and civil service Special governement distribution Private Consumers Chart 1. BULOG market operations for paddy and rice BULOG intervention private channels Source: Adapted from BULOG: The national food grain authority of Indonesia , 1982. � ��� ��� ��� ��� ���� � ��� � ��� � �� � � �� � � �� � � �� � �� � ���� � ���� Floor Price Average Producer Price �� ������������������������������������������������� ��������������������������� �����!���� "�����#��"�������$��%��� ��&�'������ �!����������� �!�� 5 capacity, Dolog task forces are prepared to buy directly from farmers. Bulog also purchases paddy or rice from private traders. The government announced floor price requires certain quality standards including moisture content, percent of broken and discolored grains, etc. If the grain quality is not met as specified, Bulog’s agents may adjust the purchase price in the field according to the prevailing price list. Figure 1 shows the historical relationship between government announced floor price and farm level paddy price from 1974 to 1993. The purpose of Bulog purchasing rice is to keep the market price near the floor price and of releasing stocks to keep the retail market price at or below the desired inflation rate. These activities are mainly supported by stock management. Since rice production in Indonesia is continuous, rice is being harvested somewhere in the country at any given time. In Java, which accounts for 60% of total production, rainy season rice is planted in November and harvested from March. The data indicate that Java rice production dominates Bulog's monthly domestic procurement, with most of Bulog’s purchases being conducted during March, April, and May. Bulog's domestic procurement of rice has never been over 10% of total rice production. Although Indonesia claims to have been self-sufficient in rice in 1985, it has been importing on and off since 1987. The amount of imports reached three million tons in 1995 following the drought year. Bulog storage facilities are scattered throughout the country and total capacity of Bulog’s warehouses is around 3.5 million tons. In recent years, Bulog maintains an average of two million tons of rice per year as a combined operational stocks, buffer stock and surplus stock. Operational stock, held for military and civil service, is 500 thousand tons a year. Buffer stock, sometimes refereed to as "food security reserve stock" to stabilize prices, is 6 around one million tons. Surplus stock is the excess of rice above operational and buffer stock. Bulog's occasional releases during the last ten years have averaged around 8% of total available rice. Fertilizer subsidies have also been an important instrument of rice policy in Indonesia. Since the late 1960s, fertilizer subsidies have been given to farmers by setting the wholesale prices of urea, triple super phosphate (TSP), and ammonia. Village cooperatives (KUD) and traders are allowed to distribute fertilizers to farmers at the official retail price level. Domestic fertilizer manufacturing plants were constructed in the mid-1970s in order to ensure adequate supplies (Pearson, et al., 1991). Timmer (1985) estimated that approximately one-half of the growth in rice production from 1968 to 1984 was attributed to improved incentives to farmers created by stable rice prices and fertilizer subsidies. However, the government has been gradually phasing out the fertilizer subsidy program and, from the beginning of 1994, only urea is being subsidized. The ratio of paddy support price and subsidized domestic urea fertilizer price increased from 1970 to 1985, and was constant during the past decade. Development of irrigation infrastructure and maintenance, transportation facilities, research and development, and dissemination of seeds and technologies for high yielding varieties are among the policy instruments used in Indonesia. Various intensification programs are known by the acronyms BIMAS, INMAS, INSUS, and SUPRA-INSUS. The first of these, the BIMAS programs, helped farmers to use improved seeds, fertilizers, pesticides, and adopt better cultivation and water management practices. The INMAS programs provided farmers with improved access to capital and extension services related to rice production. INSUS and SUPRA-INSUS programs were designed to accelerate technology adoption by requiring farmers groups of 50 to 100 with contiguous plots to make joint decisions on seeds, planting times, pest management, and off-season crop choices (Piggott, et. al., 1993). As a result, paddy production has more than doubled, from 21 million tons in 1973 to over 46 million tons in 1994. Most of these increases were due to increases in yields, from 2.56 to 4.34 tons of rice per hectare compared to an increase in area harvested from 8.4 to 10.6 million hectares in the same period (Table A.1.1.). Indonesia rice yields are among the highest in the world, although, there is still scope for quality improvement. 2.2 Soybean Soybean is one of the important secondary crops, usually grown following rice. Some argued that soybean production in Indonesia is largely inefficient mainly due to the low yield (Rosegrant, et al., 1986; Wiebe, 1990). With the five year development plan (PELITA V) 1988/89 -1993/94, a diversification program has been adopted to promote the development of soybean along with other secondary crops. Government policies to stimulate the domestic production and stabilize the retail market prices for soybean included price level policy, trade policy, and input subsidies. Earlier, credit and chemical fertilizers (urea, TSP, ammonia) were being subsidized to soybean farmers, however, these programs have been gradually Price received by farmers and millers. 4 7 phasing out and only urea is being subsidized in 1994 . Currently, Indonesia is producing nearly one and a half million tons of soybean and importing half a million tons each year to meet domestic consumption (Table A.1.2.). Average annual soybean production growth is 7%, which is largely due to increase in area harvested. Bulog is assigned to implement price and trade policy for soybean. Chart 2 illustrates the Bulog market operation for soybean since 1977. Bulog sets the import quotas to protect the domestic producer. The floor price for soybean is supported by direct purchase of KUDs, Bulog task force, and Association of Wholesalers and Distributors (AWSD) from the farmers. However, the government floor price has not been effective since the average producer price is always above the floor price (Table A.1.5.). As a result from the beginning of 1983, Bulog no longer purchases from domestic farmers. Majority of Bulog imports are distributed directly to the soybean product manufacturers. The Bulog ceiling price for manufactures and retail markets are monitored and adjusted by quantity allocations to the region. In recent years, Bulog annual soybean stock average 124 thousand tons, which is 6% of annual domestic demand. 2.3 Sugar Sugarcane is widely considered as one of the strategic commodities in Indonesia. Price level policies, trade policy, and input subsidies are also used to support domestic cane sugar production and achieve market stability. Since 1982, marketing of sugar cane and white sugar has been controlled by Bulog. The Sugar cane industry has been heavily subsidized by the government, amounting to US $313 million in 1991, with 75% in the form of price setting (Soentoro and Sudaryanto, 1996). Chart 3 displays the Bulog market operation for sugar. At the farm level, cooperatives are authorized to collect the sugarcane from farmers and send it to the mills. Farmers receive payments for their cane in terms of sugar and cash from cooperatives. The floor price support is mainly supervised by cooperatives. White sugar from mills and imports is distributed by Dolog to large industrial users, cooperatives, and private distributors in each province. The allocations to provincial distributors are increased or decreased according to how the price moves in relation to the desired ceiling price. The government has maintained the floor price for sugarcane in proportion to that of rice. The ratio of the floor price for sugarcane (ex-factory price ) to the price of rice has been4 maintained at 1.5 since 1983 (Soentoro and Sudaryanto, 1996). During the last decade, the government announced floor prices has been on average 64% above the London fob sugar price, protecting local producers from the world market (Table A.1.6.). Nevertheless, 8 Chart 2 BULOG market operation for soybean Private sector channels BULOG intervention Source: Adapted from BULOG: The national food grain authority of indonesia, 1982. Farmers Village Coops (KUDs) BULOG Task force Imports DOLOG (BULOG) Floor price Soy products manufacturess Association of whole salers and distributors Imports authorized by BULOG Retail market Private consumers Retail Ceiling price 9 Source: Adapted from BULOG: The national food grain authority of Indonesia , 1982. Private sector channels BULOG intervention Chart 3 BULOG market Intervention for white sugar Farmers Priivate sugar plantation Lampung Village Coop (KUDs) DOLOG (BULOG) Imports Sugar mills Private sectos Distributors Cooperative distributors Cooperative wholesalers Private sector wholesalerss Retail markets Private consumers Small users Large users Industrial consumers Ceiling price Floor price BULOG price control Sugar cane channel 10 domestic cane sugar productivity has been stagnant for the last ten years. While production increases over the past ten years have been nearly 5% per year, they are largely due to increases in area harvested. In 1993, Indonesia produced 2.4 million tons of sugar and imported 237 thousand tons, which is 10% of the total sugar supply. In recent years, Bulog's annual stock of sugar has been approximately 42% of domestic production. 2.4 Wheat and wheat flour All wheat consumed in Indonesia is imported. Since 1971, Bulog has regulated both international and domestic marketing of wheat and wheat flour (Chart 4). PT Bogasari in Java and PT Berdikari in Sulawesi are the only two private flour mills which received licenses from Bulog to import, process, and distribute wheat and wheat flour. Between the two, PT Bogasari maintains a dominant position, with 87% of the domestic market and, in addition, they strongly influence or control management of PT Berdikari (Kompas, 1995a). Imported wheat grains are sent directly to the private flour mills at a price determined by Bulog. Wheat flour is distributed to industries and other consumers through both private and cooperative distributors, at a price determined by Bulog. Similar to other controlled crops, the price ceiling for wheat flour is realized by adjusting the quantity distributed from the factory. The wheat grain price for flour mills has been fixed at 141 Rp/kg since 1984 regardless of price fluctuations in the world market. The difference is being subsidized by Bulog with the fund provided by processing and distribution industry. However, studies argued that the subsidy seemed to be mostly incurred to consumers (Kwik Kian Gie, 1995; Kompas 1995b). The amount of subsidy reached approximately US $ 300 million in 1994 (Kompas 1995b). 2.5 Funding and expenditure Bulog receives its funding in terms of credit from the Ministry of Finance via the Central Bank, and the amount of credit is limited by the current value of stocks in the pipeline. The interest rate charged to Bulog is adjusted periodically; the annual rate is 12% for the year 1996 (Bulog, 1996). Bulog, then makes payments for the imported commodities by opening letter of credits to the supplier, and for domestic procurement by transfer payment through the bank upon receiving complete documentation of the transactions, while farmers receive cash for their commodities from Bulog’s agents. Private traders use their own funds and KUDs are financed by the Peoples’ Bank of Indonesia with the credit limit depending upon potential availability of commodities in the region. The annual rate charged to KUD was 14% for 1996 and is also adjusted periodically (Bulog, 1996). Bulog annual expenditure for purchase, release, stock management, distribution, imports, and administration is approximately US$ 1.5 billion in 1991-92 (Bulog, 1995). 11 Private consumers Retail markets and institutional users Private industry Coop industry Private wholesalers Coop wholesalers Private sector distributors Coop distributors Flour mills BULOG buys wheat grain for Imports BULOG price control Ceiling price delivery to flour mill Bulog intervention Private sector channels Source: Adapted from BULOG: The national food grain authority of Indonesia , 1982. Chart 4 BULOG market intervention for wheat and wheat flour Appendix 3 presents the full dis-aggregated SAM as described in Table 3.2.5 12 3. The Social Accounting Matrix (SAM) for Indonesia A SAM is a system for organizing economic transactions over a defined period of time (usually a year) and is in the form of a square matrix, with column sums equaling corresponding row sums. A SAM provides a single framework that reconciles both the input- output accounts (which portrays the system of interindustry linkages in the economy) and the national income and product accounts. The SAM generalizes the input-output idea that one sectors's purchase is another sector's sale to include all transactions in the economy, not just inter-industry flows. Any flow of money from, say, a household to a productive sector (representing the purchase of that sector's output by the household), or from a household to the government (representing tax payments), is recorded in the SAM as an expenditure by some actor (a column) to some other actor as revenue (a row). The second idea embodied in the SAM, derived from national income accounting, is that income always equals expenditure. While true for the economy as a whole, the SAM requires a balance in the accounts of every actor in the economy. For example, the income from sales in the agriculture sector must equal its total expenditures on intermediate inputs, labor, imports, and capital services. Traditionally, this balance is captured in double-entry bookkeeping by the requirement that the two sides of the ledger must be equal. In the SAM, incomes appear along the rows, and expenditures down the columns; thus the budget constraints require that the row sum (income) must equal the corresponding column sum (expenditure) for every actor. The SAM also distinguishes between “activities” and “commodities,” allowing for two different effects. First, it permits more than one type of activity to produce the same commodity, thereby allowing for different production technologies. For example, small- and large-scale farmers may produce the same crop (a single “commodity”), but with different factor intensities (two or more “activities”). Second, this treatment addresses several difficult problems that arise from dealing with imports. If imports are at all competitive with domestically produced goods (which is usually the case), then domestic demand will consist of both types of goods. However, only domestic goods are exported. Separating activity accounts (or the domestic production of goods) from commodity accounts (the domestic demand for goods) enables us to portray this difference. The different accounts in the SAM outline the boundaries of an economywide model. Table 3.1 presents an aggregate SAM for Indonesia for 1990, which provides a useful representation for discussing the equations of the core CGE model, while Table 3.2 shows the level of dis-aggregation of the aggregate SAM underlying the CGE model of this paper.5 Specifying a “complete” model requires that the market, behavioral, and system relationships embodied in each account in the SAM be described in the model. The activity, commodity, See Pyatt and Round (1985) for more information on Social Accounting Matrices, and for more6 information on Social Accounting Matrices for Indonesia see Biro Pusat Statistik (BPS) (1994a), the government official agency responsible for data collection and processing at the national level. 13 and factor accounts all require the specification of market behavior (supply, demand, and clearing conditions). The households, enterprise, and government accounts embody the private and public sector budget constraints (income equal expenditure). Finally, the capital and world accounts represent the macroeconomic requirements for internal (saving equals investment) and external (exports plus capital inflows equal imports) balance. 6 Table 3.1. An Aggregate SAM For Indonesia, 1990 (BILLIONS OF 1990 RP) Expenditures or Outlays Value Added Suppliers Institutions Labor Land Capital Activity Commodity Households Enterprise Government Capital World Total (1) (2) (3) (4) (5) (6) (7) (8) (9) (12) Value Added R Labor 94027.1 94027.1 e Land 13953.5 13953.5 c Capital 90616.5 90616.5 e Suppliers i Activity 355053.2 53288.7 408341.9 p Commodity 200540.3 127330.9 15502.8 64790.0 408163.9 t Institutions s Households 94027.1 13953.5 35855.3 4616.2 242.7 5723.4 3612.6 158030.9 Enterprise 54761.2 -4272.0 50489.2 Government 9204.5 3064.9 1997.8 23059.1 -4090.1 33236.2 Capital Account 24086.0 19667.5 12010.0 9026.5 64790.0 World 50045.8 7519.9 57565.7 Total 94027.1 13953.5 90616.5 408341.9 408163.9 158030.9 50489.2 33236.2 64790.0 57565.7 Source: Biro Pusat Statistik (1994a) and (1994b) 15 Table 3.2. SAM disaggregation (Activities, Commodities, factors, and institutions) Set Elements Activities/Commodities Agricultural (13) Rice, Soybeans, Maize, Cassava, Vegetables and fruits, Other food, Rubber, Sugarcane, Coconut, Palm Oil, Other non-food, Livestock, Forestry Other (21) Fishery, Oil, Mining, Food processsing, Textiles, Paper, Fertilizer, Chemical, Petroleum refinery, Cement, Steel, Other manufacturing, Construction, Electricity-Gas-Water, Trade, Restaurant and Hotels, Transportation and Communication, Services, Public Administration, Other Services. Factors of Production Labor (10) Rural paid agriculture labor, Urban paid agriculture labor, Rural unpaid agriculture labor, Urban unpaid agriculture labor, Rural production transport equipment operator and manual labor, Urban production transport equipment operator and manual labor, Rural clerical sales and services labor, Urban clerical sales and services labor, Rural professional and managerial labor, Urban professional and managerial labor, Land Capital Institutions Households (8) Agriculture: Agricultural worker, Small farmer, Medium farmer, Large farmer Other: Rural lower level, Rural higher level, Urban lower level, Urban higher level Companies Government Rest of the World The model can easily incorporate downward-sloping demand curves for exports, endogenizing the7 world price of exports, pwe. Equations (4) and (5) are cost functions arising from first-order conditions for the CES and CET8 functions. Since CES and CET aggregation functions are linearly homogeneous, the cost functions can be replaced by the accounting identities shown (showing each price as the average of a traded price and a domestic price), as the first order conditions will be incorporated in the import demand and export supply functions presented later. 16 4. Equations of the Core CGE Model The SAM presented above provides a description of the circular flow of income in the Indonesian economy from activities to factors of production, to institutions, to commodities, and back to activities. The role of the AG-CGE model is to specify the market, behavioral, and system relationships embodied in each account of the SAM. This section presents the equations of AG-CGE that capture these relationships. First, Table 4.1 lists all the model indices, parameters, and variables of the model. Second, equations defining the price system are presented, followed by equations defining production technology, value added, and the mapping of value added into institutional income. Then equations specifying the balance between supply and demand for goods by the different agents complete the circular flow. Finally, the market clearing conditions and the macro closure rules, often referred to as system constraints that the model economy must satisfy, are presented. Some notational conventions are followed consistently. Endogenous variables are presented in upper case, while parameters and exogenous variables are always lower case or Greek letters. Indices appear as lower case subscripts, and consist of sectors (i and j), primary factors of production (f), and households (h). In a few equations, an index is replaced by a specific entry from the set. Appendix 2 to this paper presents the basic elements (sets, parameters, variables, and equations) of the model in GAMS syntax 4.1 Price Equations Table 4.2. presents the equations defining prices in the model. Equations (1) and (2) define import and export domestic prices, respectively. On the import and export side, the “small country” assumption is maintained as the world price of imports (pwm) and exports (pwe) are exogenous. Both the domestic price of imports (PM) and the domestic price of7 exports (PE) are the tariff or subsidy-inclusive world price times the exchange rate (EXR). Equation (3) defines domestic commodity prices (PDC) as the domestic activity goods price multiplied by the make matrix coefficients. Equations (4) and (5) describe the prices for the composite commodities Q and X. Q is total sectoral domestic use, which is a constant elasticity of substitution (CES) aggregation of sectoral imports (M) and domestic goods supplied to the domestic market (D). X is total sectoral output, which is a constant elasticity of transformation (CET) aggregation of goods supplied to the export market (E) and goods sold on the domestic market (D) . 8 Table 4.1. Definition of Model Indices, parameters, and Variables i, j Sectors Rice Furniture Soybeans Textiles Maize Paper Cassava Fertilizer Vegetables and fruits Chemical Other Petroleum Refinery Rubber Cement Sugarcane Steel Coconut Other manufacturing Palmoil Construction Other Electricity, gas, and water Livestock Trade Forestry Restaurants and hotels Fishery Transportation and communication Oil Services Mining Public administration Food Processing Other services iag Agricultural Sectors Rice Sugarcane Soybeans Coconut Maize Palmoil Cassava Other Vegetables and fruits Livestock Other Forestry Rubber Fishery iagn Non-agricultural Sectors (iag + iagn = i) IE Export sectors IE1 Export sectors with CET function IE2 Export sectors with no CET function IE2A Export price fixed to domestic price and exports E adjusts IE2B Export price free and exports E is fixed IED Sectors with export demand equation IEDN Sectors with no export demand equation IEN Non export sectors IM Import Sectors IMN Non Import Sectors MQRN mport rationed sectors F Factors of production Agriculture Rural Paid labor Agriculture Urban Paid labor Agriculture Rural Unpaid labor Agriculture Urban Unpaid labor Rural Production & Transprt & Manual Urban Production & Transprt & Manual Rural Clerical & Sales & Services Urban Clerical & Sales & Services Rural Prof & Tech & Supervisor Urban Prof & Tech & Supervisor Land Capital ITOP Subsidized consumption sector ITARG Target price sectors Rice IESET Non CET sectors Rice 17 Table 4.1. (cont.) Parameters A AC(i) Armington function shift parameter E ENTSAV Enterprise savings AD2(i) CES shift parameter ENTTAX Enterprise tax revenue ALPHA2(i,f) CES factor share parameter ENTTF Enterprise transfers abroad ALPHA(i,f) Cobb Douglas factor share parameter ESR Enterprise savings rate AT(i) CET function shift parameter ETR Enterprise tax rate A(i,j) Input-output coefficients EXPTAX Export subsidy payments B B(i,j) Capital composition matrix EXR Exchange rate (RP per $) C CWTS(i) Consumer price weights E(i) Exports D DELTA(i) Armington function share parameter F FBOR Government foreign borrowing DEPR(i) Depreciation rates FDSC(i,f) Factor demand by sector DSTR(i) Ratio of inventory investment to gross output FLABTF Labor transfers abroad E ECON(I) Export demand constant FSAV Net foreign savings ESR0 Enterprise savings ratio FS(f) Factor supply ETA(i) Export demand price elasticirty FXDINV Fixed capital investment ETR0 Enterprise tax rate G GDPVA Value added in market prices GDP EXRB Base exchange rate GDTOT Total volume of government consumption F FMAP(hh,f) Factors to household map GD(i) Final demand for government consumption G GAMMA(i) CET functiom share parameter GOVGDP Government to GDP ratio GLES(I) Government consumption shares GOVSAV Government savings K KSHR(i) Shares of investment by sector of destination GOVTH Government transfers to households M MAKE(i,j) Make matrix coefficients GR Government revenue P PVB(i) Base value added price H HHSAV Total household savings PWMB(i) Base import price HHTAX Household tax revenue PWM(I) World market price of imports (in dollars) I ID(i) Final demand for productive investment PWSE(i) World price of export substitutes INDTAX Indirect tax revenue PWTS(i) Price index weights INT(i) Intermediates uses PXB(i) Base output price INVEST Total investment R RHOC(i) Armington function exponent INVGDP Investment to GDP ratio RHOP(i) CES production function exponent M MINIMAND Walras law minimand RHOT(i) CET function exponent MPS(hh) Marginal propensity to save by household type S SREMIT(hh) Remittance shares M(i) Imports STRANS(hh) Government transfer shares P PC(i) Consumption price of composite goods SYENTH(hh) Share of enterprise income to households PDA(I) Domestic activity goods price SYENT(f) Enterprise shares of factor income PDC(i) Domestic commodiy goods price SYTR(hh) Share of household income transferred to other households PE(i) Domestic price of exports T TC(i) Consumption tax (+) or subsidy (-) rates PINDCON Consumer price index TE(i) Tax (+) or subsidy (-) rates on exports PINDEX Producer price index TH(hh) Household tax rate PK(i) Price of capital goods by sector of destination TM20(i) Initial values of import premium rates PM(i) Domestic price of imports TMB(i) Base tariff rate PQ(i) Price of composite good TM(i) Tariff rates on imports PREMY Premium income TXB(i) Base indirect tax PV(i) Value added price TX(i) Indirect tax rates PWE(I) World price of exports Y YMAP(hh,hh) household to households map PX(i) Average output price Q Q(i) Composite goods supply Variables R REMIT Remittances B BULOGE(i) Bulog exports REMITENT Enterprise remittances BULOGM(i) Bulog imports RGDP Real GDP BULOGP(i) Bulog purchases S SAVING Total savings BULOGS(i) Bulog sales SPC(i) Variable subsidy BULSTK(i) Bulog stocks T TARIFF Tariff revenue C CD(i) Final demand for private consumption TM2(i) Import premium CH(hh) Household consumption W WALRAS1 Slack variable for savings investment equation CONTAX Consumption tax revenue WFDIST(i,f) Factor price sectoral proportionality ratios D DA(i) Domestic activity sales WF(f) Average factor price DC(i) Domesrtic commodity sales X X(i) Domestic output DEPREC Total depreciation expenditure Y YENT Enterprise income DK(i) Volume of investment by sector of destination YFCTR(f) Factor income DST(i) Inventory investment by sector YH(hh) Household income 18 PMi ' pwmi @ (1% tmi) @EXR i0im PEi ' pwei @ (1& tei) @EXR i0ie PDC j ' ' i makeif @PDAi PQi ' PDC i @CDi % PMi @Mi Qi PXi ' PDA i @DAi% PEi @Ei Xi PCi ' PQi (1 % tci & SPCi) pcupi & PCi $ 0 i0itop PXi & pxtargi % dpxtargi $ 0 i0itarg pctargi % dpctargi & PCi $ 0 i0itarg BUL stk i ' stk o i % BULOG pur i & BULOG sal i % BULOG M i & BULOG E i i0itarg stk o i % dstk i $ BUL stk i i0itarg BUL stk i $ stk o i & dstk i i0itarg PVi ' PXi @ (1&tx i)&' j PCj @aji PKi ' ' j bji @PCj PINDEX ' ' i pwtsi @PXi PINDCON ' ' i cwtsi @PCi 19 Table 4.2. Price equations 1. Import prices ( ) 2. Export prices ( ) 3. Definition of commodity prices 4. Composite good prices net of cons. taxes 5. Average producer prices. 6. Consumption prices of composite good 7. ( ) Fertilizer price ceiling 8. ( ) Producer price target floor 9. ( ) Consumer price target ceiling 10. ( ) Bulog's Stocks 11. ( ) Upper bound on Bulog's Stocks 12. ( ) Lower bound on Bulog's Stocks 13. Value added prices net of indirect taxes 14. Composite capital good prices 15. Producer price index 16. Consumer price index For an introduction to complemintarity problems applied to economic analysis that uses GAMS9 see Rutherford (1994) or Lofgren and Sherman (1997). 20 Equations (6) through (12) provide an example of complementarity problems or variational inequalities applied to an economic model capturing specific policy aspects.9 Equation (6) and inequality (7) introduces a policy tool to maintain a ceiling on consumer prices. Equation (6) distinguishes the consumption price of a composite good (PC) and the price for composite goods (PQ) by including a consumption subsidy/tax parameter (tc) and a subsidy variable (SPC). Equation (7) imposes a ceiling on consumer prices by exogenously setting pcup – the ceiling level – as a proportion of the consumption price (PC). If thei composite price (PQ) goes up, pushing the consumption price (PC) to exceed the ceiling price level, the subsidy variable (SPC), which is initially set to zero, adjusts by assuming a positive value, and thus maintains the consumption price at a level that satisfies the inequality in (7). There is a complementary slackness relationship between SPC and PC. If the PC inequality is strict, SPC is zero. Otherwise, SPC will be positive. Inequalities (8 ) and (9) describe the producer and consumer price support scheme, respectively. In (8), producer prices (PX) are not allowed to fall below an exogenously set level determined by (dpxtarg). Similarly, consumer prices (PC) cannot exceed a pre- determined level set by (dpctarg). Equation (10) specifies Bulog's stocks being equal to intial stocks (stk ) plus the net of Bulog's domestic and international trade activities. Inequalitieso (11) and (12) set upper and lower bounds on Bulog's stock levels. For example, when stock levels are low and hit the lower bound, Bulog will experience a period of stock accumulation by purchasing from domestic and international sources. Again, there is a complementary slackness relationship between the producer-price and consumer-price inequalities and the Bulog stocking and de-stocking variables. Equation (13) defines the sectoral price of value added, or “net” price (PV), which is the output price minus unit indirect taxes (tx) and the unit cost of intermediate inputs (based on the fixed input-output coefficients, a ). The product PV@X equals sectoral value added atij factor cost, which appears as a payment by the activities account to the primary factor account in the SAM. Equation (14) gives the price (PK) of a unit of capital installed in sector i. The price is sectorally differentiated, reflecting the fact that capital used in different sectors is heterogeneous. For example, a unit of capital installed in an agricultural sector can have a different composition than a unit installed in an industrial sector (e.g., more machinery and fewer buildings in the agricultural sector compared to the industrial sector). The sectoral 21 composition of capital goods by sector of origin (that is, machinery, construction, and so on) is contained in the columns of the capital coefficients matrix, b . Since each column of thisij matrix sums to unity, PK for each sector is simply the weighted average of the unit cost of capital goods required to create a unit of capital in each investing sector. This core CGE model is static, with the economywide capital stock fixed exogenously. Within the single period, the model does generate savings, investment, and demand for capital goods. However, by assumption, these capital goods are not installed during the period, so that investment simply represents a demand category with no effect on supply in the model. Hence, the heterogeneity of capital is of limited importance in the static model, since its only effect will emerge through its impact on the sectoral structure of investment final demand. In dynamic models, the heterogeneity assumption can be very important and affect the properties of different growth paths. Equations (15) and (16) define a producer price index and a consumer price index. It is convenient to have the two indices defined for purposes of using either as a numeraire (unit of account) under different macro closure rules. 4.2 Quantity Equations Table 4.3. contains the block of quantity equations, which describe the supply side of the model. The functional forms chosen must satisfy certain restrictions of general equilibrium theory. Equations (17) to (19) define the production technology and demand for factors. Equation (17) is a constant elasticity of substitution production function, and equation (18) is the demand function for factors derived from the first order conditions for profit maximization subject to equation (17). Equation (19) defines intermediate demand as a Leontief function with fixed input-output coefficients. Equation (20) specifies the commodity-activity relationship using the “make” matrix coefficients. Equations (21) to (23) distinguish between tradable and non-tradable sectors of the economy. Equation (21) contains the CET transformation functions combining exports and domestic sales, while equation (22) is defined over a set of tradable sectors with no CET function. In the case of Indonesia, this set includes the rice sector which is controlled by Bulog operations. For sectors with no exports, the CET formulation is not needed and is replaced by equation (23). Equation (24) shows the export supply functions corresponding to equation (21), which depend on relative prices (PE/PD). Equation (25) is specific for the rice sector in Indonesia and reflects the assumption that domestic rice producers do not distinguish between local domestic prices and export prices – domestic and foreign rice are assumed to 22 Table 4.3. Quantity equations 17. CES Production functionXi ' a D i @ ' f "i,f FDSC &DP i i,f & 1 DP i 18. Where FDSCif ' Xi @ "if @PVi (a D i )D P i @WFf @wfdist if FP i FP i ' 1 DP i % 1 Demand function for primary factors (First order condition for profit maximization) 19. Total intermediate usesINTi ' ' j aji @Xj 20. Commodity/activity relationshipDAi ' ' j makeij @DCi 21. Xi ' a T i (i E DT i i % (1 & (i) D DT i i 1 DT i i0ie1 Gross domestic output as a composite good 22. Xi ' Ei % Di i0ie2 23. Xi ' Di i0ien 24. Export SupplyEi ' Di PEi (1&(i ) PDAi@(i 1 DT i & 1 25. PDAi ' PEi i0ie2a 26. World Export DemandEi ' econi PW e i pwsei &0i i0ied 27. Total Supply as a composite goodQi ' a C i *i M &DC i i % (1&*i) D &DC i i & 1 DC i i0ie2a 28. Qi ' DCi i0imn 29. M i ' Di P d i @ *i P m i (1&*i) 1 1 % DC i i0im In the AG-CGE model, rice is the only commodity which is assumed to be a perfect substitute10 with exports and imports on world markets. Trade in rice is treated specially. It is possible to weaken these strong assumptions without losing the fundamental property that11 domestic and foreign goods are imperfect substitutes. 23 be perfect substitutes. Equation (26) gives the world export demand function for sectors in which the economy is assumed to have some market power (and thereby faces a downward sloping demand curve). For the AG-CGE model, we currently assume no market power for any Indonesian exports. Equations (27) to (29) give the CES aggregation functions describing how imports and domestic products are demanded, and the corresponding import demand functions, which depend on relative prices (PD/PM). Again, equation (28) is defined over sectors with no imports (imn). Note that the production function is nested. At the top level, output is a fixed- coefficients function of real value added and intermediate inputs. Real value added is a CES function of the primary factors of production. The capital input is a fixed coefficients aggregate of capital goods, but only the aggregate is shown in the production function of equation (17). Intermediate inputs are required according to fixed input-output coefficients as specified in equation (19), and each intermediate input is a CES aggregation of imported and domestic goods. In addition, in equation (21), total domestic production (X) is supplied to domestic (D) or foreign (E) markets. These three “goods” (X, D, and E) are all distinct, with separate prices, even though they have the same sectoral classification. Imports (M) and domestic goods (D) are also distinct from their composite (Q), with separate sectoral prices. The model allows two-way trade (that is, simultaneous exports and imports) at the sectoral level, again reflecting empirical realities in developing economies. 10 One implication of this treatment of exports and imports is the partial insulation of the domestic price system from changes in world prices of sectoral substitutes. Through choice of substitution elasticities, the CET and CES functions provide a continuum of tradability at the sector level. This treatment is empirically more realistic than the extreme dichotomy between traded goods (where domestic and foreign products are perfect substitutes) and non-traded goods commonly found in analytic trade models. It also permits a richer specification of import demand than the two extremes of assuming either perfectly competitive and non-competitive imports. While flexible, the particular functional forms adopted here (CES and CET) do embody strong assumptions about separability and the absence of income effects. The ratios of exports and imports to domestic sales (E/D and M/D) at the sectoral level depend only on relative prices, and the demand for factor inputs in production does not depend on the export share.11 YFCTRf ' ' i WFf @FDSCif @WFDISTif YENT ' ' f syentf @YFCTRf % REMITENT @R % PREMY YHhh ' ' f fmaphh,f @ (1 & syentf) @YFCTRf % sremithh @ (REMIT & FLABTF) @EXR % stranshh @GOVTH % ' h ymaphh,h @ sytrh @YHh % syenthhh @ (YENT & ENTTAX & ENTSAV & ENTTF @EXR ) CHhh ' (1& thhh ) @ (1&MPShh ) @YHhh & ' h ymaphh,h @ sytrh @YHhh TARIFF ' ' i tmi @PWM i @M i @EXR i0im PREMY ' ' i tm2 i @M i @pwmi @EXR i0im CONTAX ' ' i ( tc i&SPC i ) @PQi @Qi INDTAX ' ' i tx i @PXi @Xi EXPTAX ' ' i te i @PWEi @Ei @EXR i0ie HHTAX ' ' hh thhh @YHhh DEPREC ' ' i depri @PKi @FDSCi,Capital ENTTAX ' etr @YENT ENTSAV ' esr @YENT HHSAV ' ' hh MPShh @YHhh @ (1 & thhh) GR ' TARIFF % CONTAX % INDTAX % HHTAX % FBOR @EXR % ENTTAX % EXPTAX SAVING ' HHSAV % ENTSAV % DEPREC % GOVSAV % EXR @FSAV 24 Table 4.4. Income equations 30. Factor Income 31. Capital Income 32. Single Household Income 33. Household Disposable Income 34. Tariff revenue 35. Import Premuim 36. Consumption taxes 37. Indirect taxes 38. Export subsidy payments 39. Total household taxes 40. Depreciation Expenditure 41. Total enterprise taxes 42. Total enterprise savings 43. Household savings 44. Government revenue 45. Total savings Private savings % government savings % foreign savings ' Investment The mapping schemes are used to move from factor incomes to households in CGE models. In12 applications, the mapping choice is driven by the focus of the model (i.e. models concerned with income distribution will have more elaborate mappings) or by the availability of data on household expenditure patterns. 25 4.3 Income Equations Table 4.4 presents the equations which map the flow of income from value added to institutions and ultimately to households. These equations fill out the inter-institutional entries in the SAM. Many of the entries in this part of the SAM (and the income and expenditure flows they represent) will be specific to the structure of a particular economy. The distinction between parameters and variables also becomes important. While conceivably variable, many of these items are set exogenously or determined by simple share or multiplier relationships, rather than through complex behavioral representations. Equation (30) defines factor incomes, which in turn are distributed to capital and labor households in equations (31) and (32). Then in equation (33) household disposable12 income is defined. Equations (34) to (43) determine government tariff (TARIFF), import premiums (PREMY), consumption tax (CONTAX), indirect tax (INDTAX), export tax/subsidies (EXPTAX), income tax (HHTAX), and corporate taxe (ENTTAX) revenue, while total government revenue (GR) is obtained as their sum in equation (44) plus the government foreign borrowing (FBOR). The components of savings include household savings (HHSAV) from fixed savings propensities (mps) in equation (43), corporate savings (ENTSAV) as a fixed proportion of corporate income (esr) in equation (42), financial depreciation (DEPREC) in equation (40), and government savings (GOVSAV), obtained as the difference between government revenue and consumption. Total savings (SAVING) in equation (45) includes these three domestic elements plus foreign savings in domestic currency (FSAV.EXR). Note that these income equations also embody the three major macro balances: savings-investment, the government deficit, and the current account. Firms and households save fixed proportions (depr and mps) of their incomes, enterprises save a fixed share of their income (esr), government savings is the budget surplus or deficit, and foreign savings represents the capital inflow required to balance international payments, i.e., net foreign savings. Since the model satisfies Walras' Law, the three macro balances must satisfy the identity: FXDINV ' 'i PK i @DK i ' 'i PC i @ IDi . See, for example, Dervis, de Melo, and Robinson (1982) who include an appendix about the LES13 and their application in CGEs. Note that, given the definition of PK in equation (13): 14 26 The modeler must avoid the specification of independent equations for each of these components, since without some residual category, the resulting model will be overdetermined. The range of alternative macro “closures” is discussed further below. 4.4 Expenditure Equations Table 4.5. provides equations which complete the circular flow in the economy, determining the demand for goods by the various actors. Private consumption (CD) is obtained in equation (46), a Stone-Geary linear expenditure system (LES). In equation13 (47), government demand (GD) for final goods is defined using fixed shares of aggregate real spending on goods and services (GDTOT) plus the net of Bulog's sale/purchase activities. Equation (48) determines government total expenditures including Bulog's external trade activity. Aggregate nominal fixed investment (FXDINV) is calculated in equation (49) as total investment (INVEST) minus inventory accumulation. Aggregate fixed investment is converted into real sectoral investment by sector of destination (DK) in equation (50) using fixed nominal shares (kshr), which sum to one over all sectors. Equation (51) translates investment by sector of destination into demand for capital goods by sector of origin (ID), using the capital composition matrix (b ).ij 14 4.5 Market Clearing Conditions and Macroeconomic Closure Table 4.6. contains equations defining the system constraints that the model economy must satisfy. While recognizing that the model is a general equilibrium system, with all endogenous variables jointly determined, it is nevertheless useful to think in terms of matching each of these equilibrium conditions with an “equilibrating variable.” In a competitive market economy, these equilibrium conditions correspond to market- clearing conditions, with prices adjusting to clear each market. Equation (52) states that the sectoral supply of composite commodities must equal demand, and thus defines market-clearing equilibrium in the product markets. There is also an analogous sectoral market-clearing equation for domestically produced goods sold on the domestic market (D). However, from equation (29) it is evident that the ratio of imports to domestic sales is the same for all categories of imports. Thus, at the sectoral level, specifying 15The same reasoning can be used to justify why there is no separate market-clearing condition for domestic output (X), since this involves adding exports to both sides of this adjusted market-clearing condition. 27 Table 4.5. Expenditure equations 46. Private consumptionPCi @CDi ' ' hh PCi @(i,hh %$i,hh @(CHhh&' j PCj@(j,hh ) 47. Government consumptionGDi ' gles i @GDTOT% BULOGPi &BULOGSi 48. Government savingsGR ' ' i PCi @GDi%GOVSAV%GOVTH & ' itarg BULOGEitarg EXR PWEitarg % ' itarg BULOGMitarg EXR pwmitarg 49. Total fixed investmentFXDINV ' INVEST & ' i PCi @DSTi 50. Real fixed investment byPKi @DKi ' kshr i @FXDINV sector of destination 51. Investment final demandIDi ' ' j bij@DKj by sector of origin a separate market-clearing condition for domestically produced goods sold on the domestic market amounts to multiplying through both sides of equation (52) by the ratio Di/Qi. Since, if equation (52) holds, so will this new equation in which both sides are multiplied by the same number, no separate equation is required.15 The equilibrating variables for equation (52) are sectoral prices. There are eleven prices in the model which have sectoral subscripts: pwm, PWE, PM, PDC, PDA, PE, PQ, PX, PC, PV,and PK. The world prices (pwm and PWE) are treated as exogenous. Of the remaining nine, eight appear on the left hand side of price equations, leaving PDA as the variable “free” to adjust. Equation (53) defines equilibrium in factor markets. The supplies of primary factors (fsf) are fixed exogenously. Market clearing requires that total factor demand equal supply, and the equilibrating variables are the average factor prices (WFf). In the model specified here, all primary factors are Qi ' INT i % CDi % GDi % IDi % DSTi fsf ' ' i FDSCi,f ' i pwmi @Mi% ' itarg BULOGMitarg EXR pwmitarg ' ' i PWEi @Ei %FSAV%FBOR%REMIT%ENTTF&FLABTF%REMITENT % ' itarg BULOGEitarg EXR PWEitarg SAVING ' INVEST GDPVA ' ' i PVi @Xi% INDTAX%TARIFF%CONTAX RGDP ' ' i (pvbi% txbi%pxb) @Xi % tmbi @ exrb @pwmbi @Mi GOVGDP ' ' i PCi @GDi GDPVA GOVGDP ' ' i PCi @ IDi GDPVA 28 Table 4.6. Market clearing and macro economic closures 52. Goods markets equilibrium 53. Factor markets equilibrium 54. External balance 55. Saving-Investment balance 56. Value added including indirect taxes 57. Real GDP 58. Gov't to GDP share 59. Investment to GDP share intersectorally mobile: factor demands are determined through equation (18), market clearing is achieved via changing factor prices (WF ) together with exogenous sectoral-specific parameters (wdist ). In empiricalf if applications for developing countries, however, it is common to assume that sectoral capital stocks are fixed exogenously. Fixing sectoral capital stocks means that the factor demands (FDSC ) of equation (18) arei1 fixed, so that aggregate supply and demand for capital are automatically equal, and the market clearing condition for capital in equation (53) is redundant and can be dropped. Without factor conform to some initial pattern of distortions embodied in the wfdist parameters. Thus, with fixed capital mobility, however,i sectoral rental rates will not be the same across sectors, nor can they be made to In fact, the wfdist parameters become endogenous for all but one sector. This asymmetry16 occurs because fixing capital stocks in n sectors requires n new variables to ensure that equation (18) is satisfied. Since the market clearing condition is automatically satisfied, the average return to capital (WF ) is no longer needed to clear the market, so that WF together with n-1 wfdist variables are1 1 sufficient to satisfy equation (18). In practice, it is convenient to fix WF to one, and solve for the n1 wfdist parameter. The role of the real exchange rate in this class of models has been much discussed, often in a17 very confused way. These issues have been sorted out by de Melo and Robinson (1989) and in Devarajan, Lewis, and Robinson (1993), where it is shown that these models can be seen as extensions of the “Salter-Swan” model of a small, open economy with non-tradables. 29 stocks, the wfdist parameters become endogenous.16 The remaining two equations describe macroeconomic equilibrium conditions for the balance of payments and savings-investment balance. Satisfying each of these requires the specification of the variables that will adjust to achieve equilibrium and constrain other variables by fixing them exogenously. In equation (54), the balance of payments (the balance of trade in goods and non-factor services) is represented in a simple form: foreign savings (FSAV) is the difference between total imports and total exports. With foreign savings set exogenously, the equilibrating variable for this equation is the exchange rate (EXR). Equilibrium will be achieved through movements in EXR that affect export and import prices (PM and PE) relative to domestic good prices (PDA) — in other words, by changing the relative price of tradables to nontradables, or the real exchange rate. For example, an increase in the exchange rate represents a real depreciation, so that tradable prices (PM and PE) rise relative to PD. Given the export supply and import demand functions, the result will be higher exports and lower imports. Thus, from an initial equilibrium, any fall in foreign savings will lead to a new equilibrium with a higher (depreciated) real exchange rate.17 Alternative foreign exchange market closure choices are also possible. For example, the exchange rate can be fixed, and foreign savings can adjust. The last macro closure condition in equation (55) requires that aggregate savings equal aggregate investment. The components of total savings have already been discussed: government savings is determined as the residual after government revenue is spent on fixed real government consumption (GDTOT), private savings are determined by fixed savings rates, and foreign savings (in at least one closure choice) are fixed exogenously. This specification, which is used in the AG-CGE model, corresponds to a “savings driven” model, in which aggregate investment is the endogenous sum of the separate savings components. This is often called “neoclassical” closure in the CGE literature. Recent discussions of macro closure in developing country CGE models are in Chapter 8 of18 Devarajan, Lewis and Robinson (1997), as well as Robinson (1989), Adelman and Robinson (1988), Dewatripont and Michel (1987), and Rattso (1982). The seminal article on macro closure is Sen (1963). See also Taylor (1990). 30 As with the balance of payments equation, there are alternative ways to achieve savings-investment equilibrium. Various “investment driven” closures have been used in which aggregate investment (INVEST) is fixed and some savings component or parameter (such as mps, esr, or even FSAV) becomes endogenous. “Keynesian” closures, which incorporate multiplier mechanisms, are possible as well. 18 Equations (56) and (57) define nominal and real GDP. Real GDP (RGDP). Both are defined from the value added side. They can be used to define the GDP deflator, which is sometimes chosen as the numeraire in CGE models. In the AG-CGE model, the numeraire is the consumer price index, PINDCON. With this numeraire, changes in nominal incomes measure real welfare changes and nominal wages measure real wages in consumer prices, which is convenient for purposes of presenting results. After macro closure decisions are made, careful counting of the equations and variables in the model indicates that the number of equations is one more than the number of endogenous variables. However, the core CGE model satisfies Walras' Law. Therefore, the equations defining the equilibrium conditions (Tables 4.4 and 4.5) are not although the choice has no effect on the solution of the model. all independent; any one of them can be dropped, thus equating the number of variables and equations. In practice, the savings-investment equation is most frequently dropped, 31 5. Base Solution, Policy Experiments, and Results The model uses from the 1990 SAM data to provide the benchmark for comparing the results of policy experiments. The base run of the model starts from the benchmark data for 1990, and then updates indirect tax rates and tariff rates to 1995 values. We also assume a fifteen percent wedge between world export and import prices of rice, compared to the initial domestic price, facing Bulog when it operates in world markets. This base solution provides the benchmark against which results from various experiments are compared. This section presents the base structure of the Indonesian economy, describes the policy experiments, and reports the results. 5.1 Structure of the Economy: Base Solution Table 5.1 presents base sectoral data and values of various elasticity parameters. The model is calibrated, using the SAM data and these elasticity parameters, so that the base solution replicates the input SAM. The base SAM is assumed to represent an equilibrium for the model economy, and the parameters of the model are initialized to insure that the model solution in fact replicates the SAM. In our case, we then change some parameters (indirect taxes, tariffs, and world rice prices) to update the model. The new base solution of the AG- CGE, which provides the benchmark for making comparisons, is thus an updated base, with some data from 1995. In the core AG-CGE model, constant elasticity of substitution and transformation (CES and CET) functions are used to represent production and trade aggregation functions. Elasticities of substitution between factors in production and elasticities of substitution between home-produced goods and imports are shown in Table 5.1. Sectoral elasticities of transformation of output into exports and home-used domestic output are also listed. Consumer expenditures are determined using Stone-Geary utility functions for each household (eight in all). Income and own-price elasticities of demand by households are listed in Appendix 4. Table 5.1 shows the structure of sectoral value added, output, trade, and trade ratios. The table is organized to focus on the agriculture sector as opposed to the rest of the Economy. Agriculture value added is 26.4 percent of total value added, while of that 26.4 percent, 16.2 percent is from Food crops, 3.5 percent from Other agriculture, 2.6 percent from Livestock, 1.9 percent from Forestry, and 2.1 percent from Fishery. The table also shows how value added is distributed among other non-agriculture sectors. 5.2 Policy Experiments To conduct a policy experiment, one or more policy parameters are changed from their initial base value and the model is then solved for a new equilibrium. We consider three sets of experiments where rice productivity shocks are introduced: (1) an adverse 32 Table 5.1. Structure of the Indonesian Economy, 1990, the base year for the Model Sectoral composition (Percent) Ratios Elasticities Domestic Exports / Imports / Substitution Transformation Production Value Added Output supply Exports Imports output domestic supply elasticity elasticity elasticity (VA) (X) (Q) (E) (M) (E/X) (M/Q) (rhoc) (rhot) (rhop) Agriculture 26.4 19.0 19.5 3.2 2.0 2.6 1.2 Food crops Rice 8.4 8.2 7.8 0.0 0.0 0.0 0.0 0.75 1.25 0.33 Soybeans 0.6 0.3 0.4 0.0 0.5 0.0 8.9 0.75 1.25 0.33 Maize 0.8 0.4 0.4 0.1 0.0 1.0 0.1 0.75 1.25 0.33 Cassava 1.1 0.5 0.6 0.0 0.0 0.0 0.0 0.75 1.25 0.33 Vegetables and fruits 4.2 2.1 2.5 0.0 0.2 0.1 0.6 0.75 1.25 0.33 Other 1.1 0.6 0.7 0.3 0.6 3.9 5.4 0.75 1.25 0.33 Total 16.2 12.2 12.3 0.4 1.4 Other Agriculture Rubber 0.4 0.2 0.2 0.1 0.0 4.1 0.1 0.75 1.25 0.33 Sugarcane 0.4 0.3 0.3 0.0 0.0 0.0 0.0 0.75 1.25 0.33 Coconut 0.7 0.3 0.3 0.0 0.0 0.2 0.0 0.75 1.25 0.33 Palmoil 0.5 0.3 0.2 0.6 0.0 17.5 0.0 0.75 1.25 0.33 Other 1.6 0.9 0.8 1.3 0.2 11.4 1.6 0.75 1.25 0.33 Total 3.5 2.1 1.8 2.0 0.2 Livestock 2.6 2.4 2.5 0.1 0.1 0.2 0.2 0.75 1.25 0.33 Forestry 1.9 1.0 1.2 0.2 0.3 1.3 1.6 0.75 1.25 0.33 Fishery 2.1 1.3 1.6 0.6 0.0 3.6 0.0 0.75 1.25 0.33 Non-agriculture 73.6 81.0 80.5 96.8 98.0 16.7 14.7 Oil 13.5 6.8 3.5 22.9 4.5 27.7 8.0 0.50 1.50 1.00 Mining 2.8 1.5 1.4 2.9 0.8 15.4 3.8 0.50 1.50 1.00 Food Processing 6.1 6.3 6.4 7.5 2.5 9.7 2.5 1.50 2.00 -0.33 Furniture 2.8 2.9 1.3 13.7 0.1 39.5 0.5 1.50 2.00 -0.33 Textiles 2.6 3.7 2.9 10.5 4.6 23.5 9.9 1.50 2.00 -0.33 Paper 0.7 0.9 1.0 0.6 1.1 5.5 6.8 1.50 2.00 -0.33 Fertilizer 0.5 0.8 0.7 0.9 0.5 9.5 4.6 0.50 2.00 1.00 Chemical 1.1 1.6 3.6 1.6 14.1 8.3 24.4 0.50 2.00 1.00 Petroleum Refinery 4.5 5.4 3.5 18.5 2.9 28.0 5.1 0.50 1.50 1.00 Cement 0.6 0.7 1.1 0.8 1.9 8.9 10.8 0.50 2.00 1.00 Steel 1.1 1.4 2.0 2.7 5.3 15.4 16.9 0.50 2.00 1.00 Other manufacturing 4.2 5.9 13.1 6.6 46.1 9.3 22.2 0.50 2.00 1.00 Construction 7.0 10.6 9.8 0.0 0.0 0.0 0.0 1.50 2.00 -0.33 Electricity, gas, and water 0.9 1.2 1.1 0.0 0.0 0.0 0.0 0.50 2.00 1.00 Trade -1.8 9.3 8.3 0.4 0.6 0.3 0.4 2.00 0.50 -0.50 Restaurants and hotels 4.2 4.1 3.7 2.0 2.0 4.0 3.4 1.25 0.50 -0.20 Transportation and communication 1.9 5.4 5.1 1.6 2.3 2.4 2.9 0.50 0.50 1.00 Services 9.7 5.9 5.5 3.3 4.5 4.6 5.2 1.25 0.50 -0.20 Public administration 9.6 5.2 5.1 0.5 3.3 0.8 4.1 1.25 0.50 -0.20 Other services 1.6 1.4 1.4 0.0 0.9 0.2 3.9 1.25 0.50 -0.20 Total 100.0 100.0 100.0 100.0 100.0 Note that we can specify more or less than five percent ceiling on consumer prices for rice.19 Bulog behavior is modeled by specifying different “regimes” defined by inequalities in prices20 and buffer stocks. The regime switches are modeled using a mixed complementarity programming model. Bulog's buffer stock amounts to three and half percent of the initial level of rice production. The21 Buffer stock is set exogenously, and can be varied. In fact policy experiments can be implemented to test the effect of varying Bulog stocking capacity in response to a productivity shock. 33 productivity shock, (2) a favorable productivity shock, and (3) a favorable productivity shock where Bulog does not intervene in the rice market. To simulate rice productivity changes, we change the shift parameter in the production function for rice. Such changes can be interpreted as resulting from a temporary shock (e.g., weather, drought) or a permanent change (e.g., adopting new technology). In either case, we assume that the economy adjusts to the change, achieving a new market equilibrium. For the first set of experiments, an adverse production shock, rice productivity is decreased in a series of five cumulative experiments. In each, rice productivity falls five percent, for a cumulative total of 25 percent decline in experiment 5. The second and the third set of experiments are similar, with sets of five cumulative experiments. In the first two sets of experiments, Bulog is assumed to stabilize producer and consumer prices within a plus-or-minus band of five percent. The nature of Bulog19 intervention depends on the direction of the price change. In the first set, with rice20 productivity falling (by 5 to 25 percent), there will be excess demand for rice and consumer prices will tend to rise. When the consumer price of rice hits the ceiling of the price band, Bulog intervenes by selling enough quantities of rice in the domestic market to satisfy the excess demand. Bulog first sells rice from its buffer stocks. In the model's stylization of Bulog behavior, once the buffer stock hits its lower limit, Bulog starts importing, buying rice on the international market at the prevailing spot price. The productivity increase experiments are21 symmetric. The productivity increase generates an excess supply of rice, which should cause producer prices to fall. When the producer price hits the floor value, Bulog intervenes by purchasing rice from the domestic market to maintain the market price at the floor value. As Bulog purchases rice, it first replenishes its buffer stock. When stocks are at maximum target levels, Bulog starts exporting at the spot world export price (which is assumed to be 30 percent below the spot world import price). 5.3 Results This section presents the results from the three sets of policy experiments focusing on the overall fiscal position of the government, changes in rice prices and quantities, and on selected macro aggregates. 34 Rice Productivity Decline: Experiment 1 When rice productivity declines, the consumer price of rice tends to increase, prompting Bulog intervention to maintain the price within the 5% band. Tables 5.2, 5.3, and 5.4 list the results of this policy experiment. Table 5.2 shows the effect of the productivity shock on the government accounts. Initially, when rice productivity drops by 5%, there is a decline in government expenditure, because Bulog is earning money by selling from its buffer stock. However, as rice productivity continues to decline and Bulog intervenes more, net government expenditure rises as Bulog is forced to purchase imports (at spot world prices) to maintain the buffer stock at its minimum target level. The information on Bulog purchases/sales and Bulog imports/exports indicate how Bulog is intervening in the rice market. As rice productivity declines by 5%, Bulog sales increase from zero in the base year to 0.25 billion Rp, and Bulog imports remain unchanged since sales from existing buffer stocks are sufficient to maintain the consumer price for rice within the band. However, as rice productivity falls by 10% or more, the volume of Bulog intervention in the rice market increases. Bulog sales cause buffer stocks to hit their lower limit, and Bulog starts importing. Below 10%, Bulog operations involve only increasing imports, which is reflected in the net government expenditure figures. Imports increase and the program becomes more costly. Table 5.3 gives more detail on the impact of rice productivity decline on the rice sector. The consumption price of rice (Pc) hits the price ceiling when productivity falls by 5%. Since a 5% price band on rice prices is maintained (consumer and producer prices), the percentage change in Pc from its base value remains the same with further declines in rice productivity. Price stabilization becomes more costly as rice productivity falls. Bulog has to pay for imports at fixed world prices, but their domestic price increase as the exchange rate depreciates in reaction to the increased aggregate imports. The domestic output of rice (X) falls with the productivity decline. The supply of rice (Q) falls by less, as Bulog sells stocks and imports. At the macro level, the aggregate effects of an adverse rice productivity shock, shown in Table 5.4, include a significant contraction in real GDP (-4.3% with a 25% decline in rice productivity), as rice output falls. Government consumption net of Bulog sales fall, while imports increase. The increase in real imports leads to a significant depreciation of the real exchange rate (2.8%). The depreciation is required to generate additional exports to pay for the additional imports. Both aggregate exports and imports increase. The macro impact of this scenario is significant, even though rice is a relatively small share of GDP (about 6%). Bulog operations matter at the economywide level. 35 Table 5.2. Government accounts, rice productivity decline* (BN. 1990 RP) Base Rice Productivity decline values 5% 10% 15% 20% 25% Expenditure BULOG imports / (exports) 0.00 0.00 1.41 3.04 4.70 6.37 BULOG purchases / (sales) 0.00 (0.25) (1.74) (3.16) (4.56) (5.93) Fertilizer subsidy 0.00 0.00 0.00 0.02 0.05 0.08 Government consumption 15.07 14.94 15.08 15.22 15.37 15.51 Government savings 10.24 10.35 10.79 10.92 10.99 11.02 Government transfers 5.72 5.72 5.72 5.72 5.72 5.72 Total Expenditures 31.04 30.76 31.26 31.77 32.27 32.78 Revenue Consumption tax / subsidy 0.00 0.00 0.00 -0.02 -0.05 -0.08 Enterprise tax 21.75 21.56 21.84 22.14 22.44 22.74 Foreign borrowing -4.09 -4.06 -4.12 -4.18 -4.24 -4.30 Household tax 2.02 2.00 2.00 2.00 2.00 2.00 Indirect taxes 8.25 8.17 8.43 8.69 8.95 9.21 Tariff revenue 3.11 3.09 3.10 3.11 3.12 3.14 Total Revenue 31.04 30.76 31.26 31.77 32.27 32.78 * 5% variation in producer and consumer prices is allowed 3.5 % stocking capacity for BULOG 36 Table 5.3. Rice prices and quantities, rice productivity decline Base Rice productivity decline values* 5% 10% 15% 20% 25% Percent change in prices**: Domestic price of exports (Pe) 0.85 -0.77 0.65 2.19 3.73 5.25 Domestic price of imports (Pm) 1.15 -0.77 0.65 2.19 3.73 5.25 Average output price (Px) 0.99 5.19 5.15 5.12 5.08 5.05 Price of composite good (Pq) 0.99 5.00 5.00 5.00 5.00 5.00 Domestic activity goods price (Pda) 0.99 5.19 5.16 5.12 5.08 5.05 Domestic commodity goods price (Pdc) 0.99 5.00 5.00 5.00 5.00 5.00 Consumption price of composite good (Pc) 0.99 5.00 5.00 5.00 5.00 5.00 Percent change in quantities**: Exports (E) 0.00 0.00 0.00 0.00 0.00 0.00 Imports (M) 0.01 14.01 inf*** inf inf inf Domestic output (X) 29.71 -3.79 -12.35 -20.68 -28.87 -36.91 Composite goods supply (Q) 30.61 -3.00 -6.95 -10.87 -14.71 -18.49 Domestic activity sales (DA) 29.70 -3.79 -12.35 -20.69 -28.87 -36.92 Domestic commodity sales (DC) 30.59 -3.79 -12.35 -20.69 -28.87 -36.92 * For quantities, Base values are in bn. 1990 Rp ** From base values *** inf = infinite change from zero base Table 5.4. Macro results, rice productivity decline Base Rice productivity decline values* 5% 10% 15% 20% 25% Percent change in real: GDP 209.0 -0.3 -1.3 -2.3 -3.4 -4.3 Private consumption 128.6 -0.7 -0.7 -0.8 -1.0 -1.1 Investment 55.6 0.8 -0.3 -1.4 -2.6 -3.7 Government demand 15.1 -1.6 -11.0 -20.0 -28.8 -37.5 Exports 57.4 0.0 1.8 3.7 5.7 7.6 Imports -47.7 0.0 2.1 4.5 6.9 9.2 Exchange rate** 1.7 -0.5 0.1 1.0 1.9 2.8 * Base values are in bn 1990 Rp ** The real exchange rate is defined as the nominal exchange rate deflated by the producer price index (a weighted average of prices of sold domestically with the weights being the share of each sector in the value of total domestic sales of domestic output domestic output). 37 Rice Productivity Improvement: Experiment 2 When rice productivity improves, the fall in the producer price of rice prompts Bulog intervention to maintain the 5% price band. Tables 5.5, 5.6, and 5.7 present the results of this policy experiment. Similar to the productivity decline experiment, Table 5.5 shows the impact of a favorable productivity shock in the rice market on the government accounts, Table 5.6 provides detailed results for the rice sector, and Table 5.7 lists the aggregate effects. This experiment is the reverse of the first one, but the results are not perfectly symmetrical. In this case, Bulog operations will be reversed. Instead of selling rice to reduce excess demand, Bulog will have to purchase it to reduce excess supply. Production of rice increases by 39% under a 25% increase in productivity (Table 5.6). Instead of importing rice to support its sales, Bulog will export surplus rice in excess of its stocking needs. Given that import prices of rice are much higher than export prices, when Bulog intervenes by selling rice on the world market, the export earnings are less than the corresponding import costs for the same amount of rice when Bulog imported rice in the first experiment. Table 5.5, shows how Bulog purchases and exports increase as rice productivity improves. Bulog operations lose money (see the first two rows of Table 5.5) – more than under the productivity decline scenario. To support the domestic price, Bulog purchases rice at the support price and sells at a lower price to world markets. After a 5% productivity improvement, Bulog starts exporting, which causes a real appreciation of the exchange rate and changes in the structure of production. Total government revenue falls, largely because indirect tax revenue falls. The shift in the structure of production is towards goods with lower indirect tax rates (e.g., agriculture). The result is that, with productivity increases, the government deficit increases (government savings fall in the expenditure account). The asymmetry of response between experiments 1 and 2 is shown by the exchange rate effect (Table 5.7). In the first experiment, the exchange rate depreciates by 2.8% with productivity decline of 25%, while in the second the exchange rate appreciates by only 2.5% when productivity increases 25%. The difference is due to the fact that the export price of rice is well below the import price. Increased exports generate smaller increase in earnings, and less exchange rate appreciation is required to generate the additional imports financed by the export earnings. Rice Productivity Improvement Without Bulog Intervention: Experiment 3 This experiment is the same as Experiment 2 except that there is no Bulog intervention. Prices are free to adjust to changed market conditions. Note that the 38 Table 5.5. Government accounts, rice productivity improvement* (BN. 1990 RP) Base Rice productivity improvement values 5% 10% 15% 20% 25% Expenditure BULOG imports / (exports) 0.00 0.00 (0.99) (2.20) (3.39) (4.57) BULOG purchases / (sales) 0.00 0.13 1.53 2.90 4.28 5.68 Fertilizer subsidy 0.00 0.00 0.00 0.00 0.00 0.00 Government consumption 15.07 15.21 15.09 14.97 14.84 14.72 Government savings 10.24 10.27 9.55 9.07 8.56 8.02 Government transfers 5.72 5.72 5.72 5.72 5.72 5.72 Total Expenditures 31.04 31.34 30.90 30.46 30.02 29.58 Revenue Consumption tax / subsidy 0.00 0.00 0.00 0.00 0.00 0.00 Enterprise tax 21.75 21.94 21.72 21.48 21.24 21.00 Foreign borrowing -4.09 -4.12 -4.08 -4.03 -3.98 -3.93 Household tax 2.02 2.03 2.03 2.03 2.03 2.02 Indirect taxes 8.25 8.35 8.11 7.87 7.63 7.39 Tariff revenue 3.11 3.13 3.12 3.11 3.10 3.09 Total Revenue 31.04 31.34 30.90 30.46 30.02 29.58 * 5% variation in producer and consumer prices is allowed 3.5 % stocking capacity for BULOG 39 Table 5.6. Rice prices and quantities, rice productivity improvement Base Rice productivity improvement values* 5% 10% 15% 20% 25% Percent change in prices**: Domestic price of exports (Pe) 0.85 0.78 -0.30 -1.54 -2.79 -4.04 Domestic price of imports (Pm) 1.15 0.78 -0.30 -1.54 -2.79 -4.04 Average output price (Px) 0.99 -5.00 -5.00 -5.00 -5.00 -5.00 Price of composite good (Pq) 0.99 -4.82 -4.84 -4.87 -4.90 -4.93 Domestic activity goods price (Pda) 0.99 -5.00 -5.00 -5.00 -5.00 -5.00 Domestic commodity goods price (Pdc) 0.99 -4.82 -4.85 -4.87 -4.90 -4.93 Consumption price of composite good (Pc) 0.99 -4.82 -4.84 -4.87 -4.90 -4.93 Percent change in quantities**: Exports (E) 0.00 0.00 inf*** inf inf inf Imports (M) 0.01 0.00 0.00 0.00 0.00 0.00 Domestic output (X) 30.14 3.36 12.09 20.81 29.64 38.56 Composite goods supply (Q) 31.05 3.35 12.08 20.80 29.63 38.57 Domestic activity sales (DA) 30.14 3.36 12.09 20.81 29.64 38.57 Domestic commodity sales (DC) 31.03 3.36 12.09 20.81 29.64 38.57 * For quantities, Base values are in bn. 1990 Rp ** From base values *** inf = infinite change from zero base Table 5.7. Macro results, rice productivity improvement Base Rice productivity improvement values* 5% 10% 15% 20% 25% Percent change in real: GDP 209.0 0.3 1.1 2.0 3.0 3.9 Private consumption 128.6 0.8 0.6 0.5 0.5 0.5 Investment 55.6 -0.8 0.0 1.0 2.0 3.0 Government demand 15.1 0.9 10.6 20.2 29.9 39.7 Exports 57.4 -0.0 0.5 1.1 1.8 2.5 Imports -47.7 -0.0 0.6 1.3 2.2 3.0 Exchange rate** 1.7 0.4 -0.2 -1.0 -1.7 -2.5 * Base values are in bn 1990 Rp ** The real exchange rate is defined as the nominal exchange rate deflated by the producer price index (a weighted average of prices of domestic output sold domestically with the weights being the share of each sector in the value of total domestic sales of domestic output). In fact, the domestic price falls below the export price after the third step (15% productivity22 increase). At that point, the free market should start exporting. The last two steps thus overstate the displacement of resources out of rice. 40 domestic market is assumed to absorb all the increased supply of rice. The results, focusing22 on the differences from experiment 2, are shown in Figures 5.1, 5.2, and 5.3. Figure 5.1 show what happens to agricultural and non-agricultural production. With Bulog intervention, the rice sector draws resources (capital and labor) away from other sectors, forcing more resources into agriculture than the free market would justify. For example, with a 25% increase in productivity, rice output increases by 17% (not tabulated), compared to 39% with Bulog intervention (Table 5.6). Also, without Bulog intervention, government revenue increases (not tabulated), while in the Bulog case government revenue falls. Figure 5.2 shows the change in agriculture and non-agriculture imports imports. With Bulog intervention, the exchange rate appreciates. Without Bulog intervention, there is no increase in rice exports and a slight depreciation of the exchange rate, as increased income leads to higher demand for imports. The difference is that, with Bulog intervention, all imports rise and there is displacement of domestic non-agricultural production – the Dutch disease. The same effect is seen Figure 5.3, which shows the comparative effects on exports. They mirror the import effects except that, of course, agricultural exports (which include Bulog rice exports) rise while non-agricultural exports fall. Figure 5.4 shows the differential impact of experiments 1 and 2 on the structure of agricultural production. The effect of Bulog intervention is dramatic, keeping agricultural resources in rice that would otherwise move to other crops, especially high-value crops such as fruits and vegetables. Other crops are also affected significantly. Table 5.8 compare changes in GDP deflators with and without Bulog intervention with a 25% increase in rice productivity. With base values equal to 100 and the consumer price index being our numeraire, there is no effect on consumption deflators. With Bulog intervention, consumers are relatively worse off as the deflators for all non-consumption categories fall relative to consumer goods. Without Bulog intervention, the effects are reversed. The prices of non-consumer goods rise relative to consumer goods, so consumers are much better off. Table 5.9 gives more detail on the changes in the real and nominal value added shares with a 25% rice productivity improvement with and without Bulog. Bulog operations do not allow large price changes, as evident from Table 5.8, such that the gains from the rice productivity improvement are less spread to other sectors of the economy. Without Bulog, part of the productivity gain is spread across the rest of the economy as output increase and associated productivity gain leads to lower rice prices - nominal share of rice falls while real share rises. In other words, the impact of Bulog intervention on the real share of value added is favorable only to the rice sector. Without Bulog intervention, gains from rice productivity improvement spread across the Indonesian economy. 41 Figure 5.1. Change in the value of non-agricultural production with rice productivity improvement Change in the value of agricultural production with rice productivity improvement Figure 5.2. Change in the value of non-agricultural imports with rice productivity improvement Change in the value of agricultural imports with rice productivity improvement Figure 5.3 Change in the value of non-agricultural exports with rice productivity improvement Change in the value of agricultural exports with rice productivity improvement BASE 5 10 15 20 25 331 332 333 334 B n. R p BASE 5 10 15 20 25 46.5 47 47.5 48 B n. R p BASE 5 10 15 20 25 52 53 54 55 56 B n. R p 70 75 80 85 B n. R p With Bulog intervention Without Bulog intervention Rice productivity improvement 0.8 0.9 1 1.1 1.2 1.3 B n. R p With Bulog intervention Without Bulog intervention BASE 5 10 15 20 25 1 2 3 4 5 6 7 B n. R p With Bulog Intervention Without Bulog Intervention 42 Figure 5.4. Change in the value of rice production with rice productivity improvement Change in the value of Fruit and Vegetables production with rice productivity improvement Change in the value of other agriculture production with rice productivity improvement BASE 5 10 15 20 25 7.5 8 8.5 9 B n . R p BASE 5 10 15 20 25 25 30 35 40 45 B n . R p BASE 5 10 15 20 25 30 35 B n . R p With Bulog Intervention Without Bulog Intervention 43 Table 5.8. GDP deflators with and without Bulog intervention with a 25% improvement in rice productivity GDP Deflators Base With BulogWithout Bulog Consumption 100 100 100 Investment 100 97 104 Government 100 97 105 Exports 100 96 104 Imports 100 96 104 GDP 100 99 101 Table 5.9. Changes in real and nominal value added shares with a 25% rice productivity improvement (%) Base shares (%) Shares with Bulog (%) Shares without Bulog (%) Nominal Real Nominal Real Nominal Real Agriculture Rice 6.6 6.7 8.7 9.1 5.4 7.5 Fruits and Vegetables 3.7 3.7 4.0 3.5 3.6 3.8 Other crops 5.9 5.9 6.2 5.5 5.9 6.1 Livestock 2.3 2.3 2.5 2.3 2.4 2.4 Forestry 1.7 1.7 1.5 1.6 1.7 1.7 Fishery 1.8 1.8 1.9 1.8 1.9 1.9 Consumer goods 9.4 9.5 8.8 9.0 9.6 9.5 Intermediate capital goods 22.7 22.5 21.5 21.8 22.8 22.0 Services 45.4 45.5 44.4 45.2 46.3 45.0 Total 100 100 100 100 100 100 44 6. Conclusion Indonesia has a long history of intervention in agricultural markets, especially rice. The goal of price and farm income stabilization has justified extensive intervention and the creation of Bulog, which buys and sells on the domestic market to maintain the price within a specified band and is the sole agent for buying and selling rice on international markets. Bulog also maintains buffer stocks within a specified band, and operates in the world market when necessary to achieve its target stocks, exporting or importing as necessary. Starting from an agricultural-focused computable general equilibrium (CGE) model of Indonesia, we have modeled Bulog’s behavior using a mixed complementarity approach that allows the specification of inequalities and shifts of policy regime as prices and/or stocks move within specified bands. We have used this model to explore the impact on the Indonesian economy of changes in the productivity of rice production under different assumptions about the operation of Bulog. Our empirical results support a few conclusions. Bulog operations have significant impact on government accounts and macro variables. Policy intervention in the rice market reverberates throughout the Indonesian economy, which is not surprising given that rice production accounts for about 7% of GDP (in 1990). The links between rice and the rest of agriculture, and between agricultural and non-agricultural sectors, are important. If Bulog operates to maintain the rice price when there are significant increases in rice productivity, the results are: • Rice production goes up dramatically, and the price support scheme attracts more resources into rice production. Instead of releasing resources to other high-value agricultural uses (e.g., production of fruits and vegetables), the policy draws resources away from them. The result is an inefficient allocation of resources within the agriculture sector and the rest of the economy.. • With increased rice production, Bulog operations lead to significant subsidized rice exports. The result is an appreciation of the real exchange rate, which leads to increased imports and a bias against other exports, especially of non-agricultural products. The result is an inefficient allocation of resources between agriculture and non-agriculture sectors. • The prices of non-consumer goods (intermediate and capital goods) fall relative to the prices of consumer goods, especially food. Consumers are relatively worse off. 45 • The price support program is expensive and strains the government accounts, e