You are on page 1of 20

This article was downloaded by:[Swets Content Distribution] On: 14 February 2008 Access Details: [subscription number 768307933] Publisher

: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

International Economic Journal
Publication details, including instructions for authors and subscription information:

A SAM-CGE Demonstration Model for Indonesia: Static and Dynamic Specifications and Experiments
G. A. Abbink a; M. C. Braber a; S. I. Cohen a a Erasmus University Rotterdam,

Online Publication Date: 01 September 1995 To cite this Article: Abbink, G. A., Braber, M. C. and Cohen, S. I. (1995) 'A SAM-CGE Demonstration Model for Indonesia: Static and Dynamic Specifications and Experiments', International Economic Journal, 9:3, 15 - 33 To link to this article: DOI: 10.1080/10168739500000017 URL:

PLEASE SCROLL DOWN FOR ARTICLE Full terms and conditions of use: This article maybe used for research, teaching and private study purposes. Any substantial or systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material.

Downloaded By: [Swets Content Distribution] At: 16:31 14 February 2008

Volume 9, Number 3, Autumn 1995

G. A. ABBINK. M. C. BRABER AND S. I. COHEN Erusrnus Universioj Rotterdam

In this article it will be demonstrated how a simple static Computable General Equilibrium model of the Indonesian economy can be constructed using the Social Accounting Matrix as a database. It will be shown that under a few assumptions the constant labour force and capital stock in the static model can be elaborated 'to dynamic specifications. In both static and dynamic versions the effects of productivity increases are investigated, leading to the conclusion that with respect to income and employment generating effects innovating sectors are worse off than non-innovating sectors due to low elasticities of demand. [D58,033]

The empirical application of general equilibrium models can be traced back to 1960 when oha an sen presented his multisectoral model f o r the Norwegian economy. Since the mid-seventies, the stream of Computable General Equilibrium (CGE) models has grown remarkably, which mainly can be attributed to the availability of faster and cheaper conlputers and to the renewed recognition of the importance of relative prices for the allocation of resources. CGE models proved to be very useful tools for the analysis of tax reforms, international trade, resource allocation and income distribution. The majority of CGE models are applied to developing countries, there are increasingly more applications to advanced countries as well.' computable General Equilibrium (CGE) models are especially appropriate for operation within a Social Accounting Matrix framework. A CGE model describes the whole circular flow of a market economy, while maintaining accounting consistency both at the macro level as on the level of individual actors. A Social Accounting Matrix (SAM) is a statistical representation of that circular flow, in which every transaction between the actors distinguished in the system is accounted for. Another common feature between CGE models and the SAM is the emphasis on disaggregation. Total production in the SAM is disaggregated by several production accounts, while CGE models distinguish between several types of producers. In the same way in the SAM consunlption is disaggregated by several household accounts which can serve to define the different types of consumers in a CGE model. The different types of labour and capital distinguished in the SAM factor accounts can be 'For a survey for developing countries see Decaluwe and Martens (1988).

The services sector. I. Finally in section 7 some concluding remarks will be made. A. In section 5 several dynamic aspects will be introduced in the model.' Under certain assumptions (Cobb-Douglas production. BRABER AND S. A good description of the way in which a CGE model can be linked to the SAM is given in Robinson (1989) 5 e e Thorbecke ( I 990) for a discussion of Indonesian CGE models. The manufacturing sector is rather underdeveloped in comparison with countries at a similar level of development. Agriculture is still the largest sector and employs the majority of the rapidly increasing labour force. It is also possible to include separate want accounts in the SAM if one wants to introduce goods and services by type of want in the CGE model. Two demonstrative simulations will be discussed in section 4. M. et al. 2. in the form of oil. is the main employer of the high-skilled labour force. As a result the mining sector takes in an important place in the industrial structure. The CGE approach offers another opportunity to transform a SAM into an econonly-wide model without using any supplementary data. which is mainly low-skilled. COHEN used to define the different types of labour and capital in a CGE model. (1982) for a comparative discuss~onof CGE models and SAM-based planning models. A more detailed description of this method can be found in Cohen (1988). .16 Downloaded By: [Swets Content Distribution] At: 16:31 14 February 2008 G. resulting in the familiar multiplier equation for the vector of endogenous variables: y = (I . utility and aggregation functions) the SAM provides all data necessary to construct a simple CGE model. If put together. The applications are demonstrated for Indonesia. a country for which elaborate SAMs are now available and for which several CGE models have been applied for more specific uses4 The Indonesian economy is characterized by the availability of rich natural resources. the mining and manufacturing sector is characterized by a very capital intensive technology and is dominated by large private and public firms. 'A large body of literature is available on SAM-based CGE models. ABBINK. The derivation of the parameters from the SAM and their calculated values are discussed in section 3.A)" x. A SAM can be transformed into an economy-wide model by subdividing the SAM into endogenous and exogenous accounts and expressing the transactions between endogenous accounts in coefficient form. C. The objective of this paper is to demonstrate the basic links between the primary form of CGE moclels and the SAM and to show how under a few assun~ptions the static CGE model can be elaborated to a dynamic one.' The flexibility of the employed framework will be demonstrated in terms of analyzing policy simulations in the static and dynamic forms of the model. as will be done in this article. THE BASIC MODEL 'See Dervis. In section 6 simulations are repeated in the context of' the dynamic model. of which the revenues are mainly accruing to government and higher income groups. which includes government services. The CGE model will be presented in section 2. The paper is organized as follows.

in which case supplementary data are required to estimate additional parameters. g and h stand for institutions and q for types of labour. At a first level total output is derived from intermediate deliveries and an aggregate of labour and capital. while the (fixed) share of real value added is put in the denominator to relate real value added to real gross output. there is no need to separate price and quantity components explicitly. which is given in (1). Every sector is assumed to act as a single representative producer. which contains only a minimal number of equations necessary to describe the economic process. The indices i and j are used to represent sectors. anlong which the exogenous ones have overbars.e. as shown in (3) Equation (4) shows that the remuneration of capital is the remainder of nominal value added after labour is paid its share. again in the short run. Total nominal value addeds by sector can easily be derived by subtracting the cost of intermediate deliveries (including those which are imported) and indirect taxes from the production value. so the volume of sectoral real value added and output can increase only through an increase in sectoral employment. to arrive at gross output in nomlnal terms.SAM-CGE DEMONSTRATION MODEL Downloaded By: [Swets Content Distribution] At: 16:31 14 February 2008 17 In this section we present the core of our general equilibrium model. In ( I ) the function for real value added is given in the numerator. . At a second level smooth substitution possibilities exist between capital and q categories of labour. The basic model is presented in Table 1. remuneration rates of capital may differ among sectors. As a result. in the short run they will hire employees until wages equal the value of their marginal product. Here. Equation ( 5 ) states that the labour market is characterized by full employment by skill level. Because nominal value added can be included in the factor demand equations (3) and (41. The real value added aggregate follows from a constant returns to scale Cobb-Douglas function in labour (distinguished by skill level) and capital. Producers are assumed to maximize profits under perfect competition in product and labour markets. For more specialized applications the model needs to be elaborated. resulting from the Cobb-Douglas technology assumption. no substitution is possible between intermediate deliveries and real value added. remuneration rates have to adjust to equalize them to the value of the marginal product of capital. whose production possibilities are represented by a two-level production function. Wages carry the burden of adjustment. The sectoral capital stock is assumed fixed for the current period. the index f is used to represent actors abroad. Parameters are represented by Greek characters. It is a general purpose model. as given in (2). It has the advantage that all parameters needed may be principally derived from the SAM. which can be called real value added. This term is multiplied by the sector price P. Consequently. Because the volume of capital is fixed. Variables are represented by capitals. 'Nominal value added incorporates besides real value added also a price component which can be called value added price. we assume Leontiev technology. while the supply of labour by skill type is fixed in the short run. i.

The Static Model ~ . C. . .0 I . j 7. List of Notations ..distribution shares T . 11 Wl.Lq i ' + n.- ..function V = value added W = wage rate X = value of gross output Y = dixposablc income A = total income L M = labour = competitive supply (quantity) imports (quantity) x = capital inc.G.. M. R j K ../1 Z h - producing sector P = 1. Endogenous variables: C = consuniption expenditure 1 = investment expenditure L = sectoral labour (quantity) P = price index R = remuneration rate of capital Exogenous variables E = cxports (quantity) K = capital (quantity) Cocfficients: a = inputloutput coefficients cp = indirect tax rates y = consumption budget shares h = labour elasticities 01 production 19 = calibration constant prod. . = 1w. - . . COHEN Downloaded By: [Swets Content Distribution] At: 16:31 14 February 2008 Table 1.~1 Zh + x rl. - r.- -~ Z l .= transfer rates o = labour inc.3Zs + x q j X j c j i f h = government 6 1 = Z. ABBINK. . . .-. A... BRABER AND S.. distribution shares p = import shares ..... - - --- Equations . 1.

. its cells can be viewed as realizations of the variables of the model. from the other equations. Quantities of exports and competitive imports are fixed. In Tablc 2 we distinguish between tlirec factor of production accounts. a capitlil account. Assuming that the SAM represents a benchmark equilibrium. foils production activities accounts and a rcst of the world (ROW) account. In equation (10) tlie price of agric~~ltilrechosen as for households. which is not an unusual assumption. To distinguish between quantity 'Although it may seem unrealistic to assume the snnie behnviour for government . corporate business and government). Equation (9) gives the equilibrium condition for the product market. In Tablc 2 the SAM for Indonesia for 1975. PARAMETERIZATION O F THE MODEL A Social Accounting Matrix gives a consistent representation of the money flows which are the nominal counterpart of all real transactions bctween the agents in an economy. Because we are primarily interested in domestic transactions we kept the relations with tlie rest of the world as simple as possible. which maximize a Cobb-Douglas utility function given a budget constraint. We assume that investment goods are supplied by one capital goods producing sector only.e.SAM-CGE DEMONSTRATION MODEL Downloaded By: [Swets Content Distribution] At: 16:31 14 February 2008 19 Factor remunerations are distributed over the institutions in cqilation (6). six household groups. but only that incomc has increased relative to the price of agriculture. We assumed a 'classical' closure70f the model. The first order condition of utility maximization implies the fixed budget shares in equation (8)". Foreign prices are assunicd to be equal to domestic prices to keep the model homogeneous in all priccs. In equation (7) transfers paid are subtracted from total institutional incomes to arrive at disposable inconie. eight institutional accounts (i. The SAM is an adapted version of tlie SAM published by Biro Pusat Statistik Indonesia (1982). 3. 'See for a discussion on closure Dewatripont and M~chcl (1987). is shown. Therefore. Because our model can only determine relative prices we have to apply a price normalization rule. Prices carry the burden of adjustment. so is all other price and nominal variables arc defined relative to agriculture. which scrvcs a the database for our CGE model. one should be aware that an increase i n disposable incomc does not imply that incomc has increased in an absolute sense. Total institutional income includes besides factor income also inconie transfers received from other institutions. so an equation for investment can be omitted because total investment follows as a residual. as stated in equation (9a). A variation on equation (6) is included as equation (6a) to allow for the fact that the total incollie of governnient coniprises also indirect taxes. in this context government behnviour conies down to fixed expenditure shares. Institutions are assumed to be groups of representative consumers. Inconle transfers are assumed to be fixed proportions of total institutional income.

COMP 12.URLO IO.TRAD I7.- - - 18 ROW TOTAL -. I4. .sirr (1982) Source: Riro Pusc~t ..FARM 6.ROW .URUP 9. CAPA 1 I5. -.--- 1 2 3 4 5 6 7 8 9 LOWN HlSK CAP1 L OWN FARM NFUP NFLO URUP URL( - .NFLO 8. . Smri.. - 326 652 38 10 59 ' . . 2 -39 .Downloaded By: [Swets Content Distribution] At: 16:31 14 February 2008 Table 2 Social Accounting Matrix for Indonesia. - - --- - 493 2422 467 351 68 268 61 268 53 - . - . .- 71 1 -- TOTAL 4163 1063 8097 373 1 ~ 615 - 642 1412 1684 1682 -- -.INDU I6.SEKV 14 2 39 332 -- 13 2 3 73 40 229 - 2 47 26 XX 322 745 545 84 .AGRI I .srik 111dorle. 1975 .NFUP 7. ITAX 13. 5 49 346 105 61 60 - 6 22 . - - - 4163 1063 8097 373 1 615 642 1412 1684 1682 5. 882 196 I60 I31 - 534 698 351 301 221 298 233 222 18.

. .C. (1 . Social Accounting Matrix Expressed in CGE Model Variables 1-3 4-1 1 FACTORS INSTITUTIONS 12 INDIRECT TAXES ACCOUNT 14-17 18 ROW TOTAL FACTORS a.. v. d. INDIRECT TAXES CAPITAL 1 fqwqqfjR. .)Vj SUBTOTAL FACTORS INSTITUTIONS .E.. .... residual residual I I ROW TOTAL %zh ..Downloaded By: [Swets Content Distribution] At: 16:31 14 February 2008 Table 3..

. The institutions are distinguished by six h o ~ ~ s c h o l d corporate business (10) and governmcnt (1 1 ) . For government an additional account is includcd to record indirect taxes ( 1 2 ) . rural non-agricultural upper class households (7). rural non-agricultural lower clash households (6). which in rows 14-17 are related to disposable income Y.(ecluation (7)). Gross operating surplus is assume to be the remuneration of the l21ctor capital (3).. In the column income transfers are recorded as outgoings of institiltions.. The three factors of production are low skilled manpower.Eqrq. highskilled manpower and capital. payments o f indirect taxes to the . The parameters (ofq and x. groups (accounts 4 to 9). managers and noncivilians.17) and by the rest of the world (non-competitive imports. The sources of investable funcls. production workers. urban lower class households (8) and urban upper class households (9). The production activities or sectors are disaggergatcd by four sectors: agriculture (14). COHEN and nominal variables prices and wages can be expressed as indices. arc placed in the row. ABBINK. mining) and construction (IS). Other outgoings are consumption expenditures. A.. savings by institutions (columns 4-1 I ) and net savings from abroad (column 18). sales and service workers are classified as low-skilled labour (account I in table 2)... 0 1 lotal income Z. row 18). T h e outgoings of the production activities are payments for intermediate deliveries by other domestic scctors (rows 14. which are all unity in the benchmark equilibrium. Disposable income is dcl'incd as total income after payment of income transfers: ( I . Agricultural labourers. industry (incl.. using constant budget shares x. trade L I I I tr. some institutions also pay a fraction r.. clerical. products) arc expressed in units which cannot be converted to well-defined physical units as tons or barrels. BRABEK A N D S. investnlent goods are assumed to be supplied by the sector industry and construction (row 15) only. This implies that the quantity variables in the model (factor services. The capital account (1 3) has investment expenditure in the column.T.i~isport(16) and governmcnt and ~ services (17). The factor high-skilled labour (2) consists of prof'essionals. Hoi~seholdsare classified according to the status 01' the head of the housc.(equation (8)). 1. An additional row is included for the (sub) total of sectoral value added. represent the share of factor payments flowing to abroad. farm workers (51. Factors receive income payments from production activities. Savings can be calculated as a residual. Incomings by institution are the institutions' share of factor income and income transfers from uthcr institutions as specified in equation (6). M...)Z. C. of their income to institutions abroad. which in Table 3 are related to the factor demand equations stated in equations (3) and (4). Product units (units of factor services) are then dcfincd as the amount of a good (factor) that in equilibrium gives a reward of one currency unit in any of its possible uses..liold into landowners (4).22 Downloaded By: [Swets Content Distribution] At: 16:31 14 February 2008 G. they are calcillated as constant fractions r. In the colunin factor incomes are distributcd in fixed proportions to domestic institutions and the rest of the world.. In Table 3 we expressed the SAM as a set of relations between the variables include in the model presented in Table I as an illustration of thc relationship between the SAM and the CGE model.

It can be seen in table 3 that the activity columns incorporate the definition of nominal value added (equation (2)). consumption goods delivered to institutions (columns 4-1 l).SAM-CGE DEMONSTRATION MODEL Downloaded By: [Swets Content Distribution] At: 16:31 14 February 2008 23 government (row 12) and factor payments (rows 1-3). It has been found that agriculture has the largest backward linkages. exported goods minus competitive imports (column 18) and investment goods (column 13). as described above. while the linkages of services are relatively low. The labour elasticities A(li can be calculated by dividing the Sactor payments in the activity columns by the value added subtotal. As the capital account of the balance of payments mirrors the current account. The incomings comprise receipts for intermediate goods delivered to other sectors (columns 14-17). is calculated by rewriting equation (1) as: By defining prices and the low-skilled wage rate as indices. The main features oS the derived parameters will be discussed briefly. X j / Pi . this cell also gives the net inflow of foreign capital. Because price indices are equal to 1 in the benchmark equilibrium. The inputloutput parameters aii' pi and cpi can be calculated by simply dividing the appropriate rows in the production activities column by their column totals. The balance of current transactions with the ROW is given at the intersection of the ROW column with the capital account row. which record the incomings of the activities. The calibration constant in the production function p. Sectoral demand as stated in equation (9) can be found in the activity rows. the parameter values can be found in the appendix. Elasticities for low-skilled labour are highest in agriculture and trade. while industry and construction is the most capital intensive sector. services have relatively high high-skilled labour elasticities. only in industry and construction the share of imports is significant: 0. In the row the payments to the ROW can be found (which are receipts from the point of view of the ROW) and in the column the receipts from the ROW are given. The ROW account (18) records the transactions with the rest of the world. The amount of imports and indirect taxes per unit of final product is small. The fact that all price variables are expressed as indices makes it easy to calculate parameters from the SAM for relations between quantity variables as well. the price ratio Pi/ Pj can be omitted in the SAM.159. Table 3 is also very useful in demonstrating how the parameters can be derived from benchmark values of nominal variables.

is equal to \ J . STATIC EXPERIMENTS In this section we demonstrate the functioning of the model by simulating two supply impulses to economic growth. / 4 * Lzl. it is assumed that a unit of high-skilled labour earns four times the wage of a lowskilled worker in the benchmark period.. w. T h e parameters which describe t h e distribution of factor income o v e r households. and L. in the dynamic niodel low-skilled and high-skilled labour have to add up to total labour supply.?V. Recipients ol' capital income arc niainly landowners and firms.)Z.. The government spends ncarly half oC its income on services. Thus high-skilled labour has to be expressed in the same units as low-skilled labour. The largest share of income is spent on agricultural products. I. The way in which this is done will be treated in detail in section 5: here we will give only the benchmark values of E l . Transfers are dominated by tlows of funds from firms to households and governnient. which is a reasonablc assunlption in the Indonesian context.' are calculated . The outcomes of thcse experiments are of a Wote that according to equation (3) in the benchmark equilibriunl \v...and n.. by the respective column totals Z. In a similar way. is equal to W . on which rural households spend even more than half of their income.Ex%. for which transfer rates are 0.24 Downloaded By: [Swets Content Distribution] At: 16:31 14 February 2008 G.. Z. but for reasons of dynaniization we used a different method. "Total income 1s defined as primary income plus income transfers received .. BRABER A N D S. ABBINK. COHEN is equal to X. 6894.' For highskilled labour and capital the same approach could be chosen in the static niodel.377 respectively. *L.22 for agriculture.. C.... On the basis of the calculated parameters it can be c o n c l u d e d that low-skilled l a b o u r income a c c r u e s niainly to low income households N ~ L I landowners. These are 10541. so capital has to be expressed in the same physical units as investment. by dividing the income transfer matrix T ~ . r.. nearly 3 for services and slightly above one half for industry and construction. To obtain this result.12 for trade and transport and 3 1 12... 13552. is equal to L.25 for services. M . A.122 and 0. High-skillcd labour income goes mainly to nonagricultural upper income groups.. which include the activities of civil servants. 4. be calculated by dividing the cells in the factor account can colunins by the column totals. In the dynamic model new investment has to be added to the existing capital stock. which are both given in the SAM.. which can be found in table 3. This implies that in the benchmark L..40 for industry and construction. Budget shares are calculated by dividing consunlption expenditures in the institutions columns by disposable income or (1 . The resulting calibration constants are close to unity for agriculture and trade and transport. The transfer rates the ratio of income transfers to total income. *L.

because of the declined value of the marginal product of capital. competitive imports). As a consequence. Experiment 1: 10% productivity increase in industry and construction.65% for landowners to -0.5% increase in real GDP (average increase in the volume of value added). 0. The change in factor incomes is not large enough to bring about a large change in income distribution: there is a slight decrease in income for all groups varying from -0.01 % price decrease." the favourable volume-effect is reversed by a -10. exports. S o all factors employed in the sector with the increased productivity undergo a decline in their income. However. As in the above experiment the efficiency parameter of the production "'Only the main results will be shown here. The fall in nominal income does not mean that welfare has decreased. The result is a 8. but should be cautiously interpreted as being specific to Indonesia. resulting in a decrease of value added in industry in relation to other sectors.35%. total remuneration of capital falls with value added in industry.42% versus 0. because very little high-skilled labour is employed in industry.1 and 0.14% for farm workers. The wage rate of high-skilled labour increases slightly relative to the low-skilled wage rate. "Aggregate demand consists of demand categories with a (negative) unitary (consumption and investment) and with a zero elasticity of demand (intermediate deliveries.78% increase in the volume of output. The welfare increase of the different groups is confirmed by a +2.SAM-CGE DEMONSTRATION MODEL Downloaded By: [Swets Content Distribution] At: 16:31 14 February 2008 25 general nature based on the restrictive assumptions of the model. tables containing elaborate simulation output can be obtained from the authors. while the owners of capital see their remuneration reduced by 5. The two experiments are: (1) 10% productivity increase in industry and construction. The increased productivity makes more than five percent of the labour employed in industry redundant. Because the consumer price index has decreased by more than the fall in nominal income."' The efficiency parameter I? of the production function in industry is increased by 10 percent.21 %. . and (2) 10% productivity increase in services. purchasing power has increased for all household groups. They can give an insight in the implications of the various assumptions and form a guidance to future elaborations of the model. Experiment 2: 10% productivity increase in services. Because a relatively large share of the national capital stock is installed in industry. the aggregate demand elasticity lies between . however. because of the generality of the assumptions and the absence of qualitative structural characteristics of the Indonesian economy in the CGE model. because demand is inelastic.

The resulting production increase in the service sector is +9.2%. although the fall is less than in the above experiment. THE DYNAMIC MODEL The model can be made dynamic by upclating the variables which are exogenous in the current year. for investment to be added to capital stock. historical growth rates or just exogenous fixation.45% respectively). But as the consumer price index has decreased by even more. so the increase in real GDP reaches all households. these groups consist for a large part of high-skilled workers. When we want to update capital stock we encounter two problems that cannot be solved by using the SAM only. However. The resulting increase in real GDP is less than one half of the increase in the industry-experiment: +1. welfare of these income groups has increased also. Again. change as a result of differences in sectoral profit rates. C.19%. In later years sectoral investment shares. The updating can be based on behavioural relationships. T h e updating of the capital stock can be described as follows. First. both variables have to be expressed in the same unit. the value added of the sector with the productivity increase falls relative to the sectors without productivity increase (-2. Here the updating will be restricted to the two factors of production: capital and the labour force. BRABEli AND S. this iluplies that national capital stock is defined as ten times benchmark gross investment. T h e decrease in the remuneration f o r t h e high-skilled leads to a worsening of the income position for non-farm upper income groups (both rural and urban: -0. Both problenls are solvable by simply assuming that capital stock grows in every sector at the same rate (5%) in the benchmark ycnr. investment is distributed over the sectors in the same proportions as the operating surplus. Investment of the previous period is added at the end of that year to the capital . Because the majority of high-skilled workers is employed in the service sector. the decrease in value added is detrimental for their relative wage rate.17%. COHEN function is increased by 10 percent. However the SAM does not give such information. Because it is also assumed that capital depreciates at 5% a year. as a result of inelastic demand. If i t is also assumed that all rates of return to capital are equal i n the benchmark year.52%). ABBINK. to keep the effects of the dynamization on the simulation results traceable. 2 1 % . M. which is only slightly higher than the increase in industry production in the above simuliition. in which the increase in production does not exceed 0. This is due partly to the fact that in Indonesia the service sector is rather small and partly to the lack of backward linkages. A . it decreases with 0 . Second. in this simulation there are hardly any spill over effects to other sectors.58% and -0. 5. there is a need to know how investment is allocated by destination in the benchmark period. Changes in exports and imports and in parameters are easily introduced. and as a result sectoral growth rates.26 Downloaded By: [Swets Content Distribution] At: 16:31 14 February 2008 C. I.

their relative prices should decline and so should their profit rates.3974 for industry and construction. which is assumed to be equal to the amount of invested capital. The denominator gives the replacement value of the capital stock. The invtstment goods are distributed over using sectors in proportions which are a function of the ratio of sectoral profitability to average profitability PRj / A P R . 0. .SAM-CGE DEMONSTRATION MODEL Downloaded By: [Swets Content Distribution] At: 16:31 14 February 2008 27 stock of the current period. : c . which happens to be the sector of industry and construction.. with k being the index of the capital producing sector in the economy. are defined in (12) as the return to invested (or financial) capital and has to be distinguished from the value of the marginal product of (physical) capital R. 0.2022 for trade and 0.3091 for agriculture. PR. In this way profits above or below the average will be eliminated in the long term through competitive investment in a competitive market. The sectoral investment shares 5: are defined in equation (1 1): sj = (3)" APR A PR The benchmark investment shares which are calculated from the sectoral shares in total operating surplus: R*Kj / 4 R*K. In the numerator R. C..jK. The average profit rate follows from ( 1 3): APR = Ki xi--* P. ' . are 0.0913 for services. Their capital stock and their production should increase faster than for other sectors.P. Capital is valued at price PI. Sectoral profit rates PR. Kj is just the operating surplus and the second term is financial depreciation. Sectors which show high profitability in the current period will increase their share in total investment.

The distribution of labour over both skill types changes endogenously. Note that the investment outlays have to be divided by the price of investment goods to determine the investment in physical units. is defined as the sum of low-skilled and highskilled labour: The growth rate of the total labour force is exogenous: We assumed the growth rate g~ to be 0.. M. BRABER A N D S.. COHEN In (14) total investment expenditure is distributed over the sectors of destination In (15) new investment is added to the capital stock and depreciation is subtracted. The second variable in which dynamics are introduced is the supply of labour: both the level of the total labour force and its distribution over the skill types change over time.. The total labour force L. The ratio between the two types can be written for the long term as follows: . ABBINK.28 Downloaded By: [Swets Content Distribution] At: 16:31 14 February 2008 G. C.025. I. A. The workers' choice between low-skilled and high-skilled employment depends on the relative wages and relative education costs.

The sectors which employ mainly low-skilled workers.12% and 3. we assume in this paper that the students cost of higher education is four times the cost of lower education. the calibration constant can be calculated from the benchmark as L. which take into account the effects for 1975 only. In the short term the low-skilled/ high skilled ratio is assumed to react with a distributed lag to changes in the wage rate: The adjustment parameter 8 is assumed to be equal to 0. The base time path is the sequence of temporary equilibria under unchanged policies. =15. DYNAMIC EXPERIMENTS With the dynamic model we simulated the basic time path from 1975 to 1980.25. To keep things simple.89%).664. represents the cost of education of type q and 5. they quit the education system and start supplying themselves as low-skilled labour. The supply of high-skilled labour . The results of static simulations. Because potential suppliers of highskilled labour are discouraged by the lower wage. the share of costs of education q charged to the user. E is the supply elasticity and K is just a calibration constant. The growth of GDP is 4. To get an idea of changes that take place over time in the base time path we will firstly discuss the rates of growth between 1979 and 1980.78% respectively. its real value added grows at 5. which can be traced back to the deteriorated position of low-skilled labour in the factor market. The largest difference appears in the decline of the wage rate of the high-skilled.37%.73%)./L. The distribution of income changes.40% per annum.27% in 1980. agriculture and trade. grow at a lower rate than other sectors: 4. The good performance of industry is mainly the result of the increased demand for investment goods (+6. and of the dynamic simulations are rather similar in the case of the productivity increase in services (experiment 2). of which industry is the only supplier. With the dynamic model we also repeated the simulations of section 4 starting in 1975. The largest contribution to this growth comes from the sector industry.SAM-CGE DEMONSTRATION MODEL Downloaded By: [Swets Content Distribution] At: 16:31 14 February 2008 where D. This leads to higher profit rates for industry and a faster accumulation of capital in this sector (+5. As a result of increased supply. although only slightly. By introducing the costs of education in the labour supply equation it is possible to analyze the consequences of changes in the costs of education charged to the student on economic welfare. 6. the price of industry has already begun to decline by 0. in favour of the higher income groups. The substitution elasticity between high-skilled and low-skilled labour is assumed equal to 1.

In that case more investment goods can be bought for a given amount of money. Therefore capital is accumulated faster than in the base time path. This can be attributed to both deniand and supply factors. M.14%. However. Although the results of static and dynaniic sinlulations differ in magnitude the main conclusion remains the same. although the decrease in value addcd is reduced in comparison with the static simulation.59%.5% in the static simulation. COHEN drops by -0.1 % as compared to 2. This phenomenon has been . also in the dynamic siniulation the effect of an increase in industrial productivity leads to a reduced iniportancc of industry and construction in the dornestic economy. This can be explained by the decline in the price index of the industrial sector by 9. So. because the low profit rate leads to a decrease in investment in this sector. The capital accumulation-effect can lead to niore significant differences between static and dynamic simulations if tlie price of capital changes substantially. We also showed that untler a few assumptions this static model can be elaborated to a dynamic one. The dynaruic elcnients in the current paper were restricted to labour supply and capital accuniulntion only. This effect has been reinforced by the fact that an increased inflow of foreign savings puslies up total investment expenditure by 5. which makes it possible to buy more capital goods at a given level of savings. To a large extent this can be explained by the fact that labour market adjustment takes place through the education system and thus reacts with a lag. value added generated in industry and construction is still lower than in other sectors. the reduced supply has a favourable effect on value added. Future analysis would require morc flexible specification of labour supply responses to wage changes in the medium term. CONCLUDING REMARKS It has been demonstrated how a simple static CGE model can be constructed using a one period SAM as database. adjustment can take place more directly if actors start supplying niore hours if their wage rate increases or i f non-participants decide to supply their skills on the labour market in reaction to a wage increasc. A. The dynaniization of labour supply proved to have only a limited effect on the simulation results. which in comparison to the static simulation reduces the wage decrease needed to adjust to the decreased demand with a full percentage point to -0. In experiment 1 substantial differences are found between the results of dynamic and static simulations. BRABER AND S. C. namely that a productivity increase leaves the innovating sector worse off in compai-ison with the non-innovating sectors. However. Cornbined with the inelasticity of deniand. Demand for the sector's output has increased resulting from the increase in total investment. The performance of the sector industry and construction has improved as compared to the static simulation. The effect on GDP is niore than twice as large in the dynamic simulation: 6. 7.83%. ABBINK.30 Downloaded By: [Swets Content Distribution] At: 16:31 14 February 2008 G. I .74%. Thus capital accun~ulationis lower than in other sectors and so is the growth of output. On the other hand thc supply of tlie sector is reduced.

2. Input-Output Parameters INDU 2 .0458 0.0637 Table A.0370 0.. Such type of micro-simulation models. In the Indonesian context.1593 0. Production Function Parameters AGRI I INDU 2 TRAD 3 SERV 4 .07 19 0.0 179 0. which resulted from consumption determined by constant budget shares and from fixed quantities of exports and imports. where the modern large-scale enterprises are the innovators.0181 0.0729 0. who found that in traditional industries the replacement of old-fashioned technologies by more efficient techniques led to a substantial decrease in employment due to demand constraints. This outcome can be explained by the low demand elasticities.0099 0.SAM-CGE DEMONSTRATION MODEL Downloaded By: [Swets Content Distribution] At: 16:31 14 February 2008 31 observed for the period 1975-1980 by Poot et al. this implies that small-scale enterprises will ~ ~ n d e r g o a deterioration in their terms of trade. as well as other differentiations between firms within a sector will be very important refinements. APPENDIX Table A. Modelling the distinction between innovating and non-innovating firms within a specific sector.1709 0. However if some firms are innovators while others are not.0 124 0. - .1. which implies that all firms are innovators or none of them are. (1984). Another reason why innovating entrepreneurs are worse off than less innovating entrepreneurs is the fact that we assumed that all firms in a sector can be described by one representative firm. Future analysis would require more flexible demand specifications to be used in combination with elasticities obtained by empirical estimation.096 1 0.--- TRAD 3 SERV 4 0. would imply a departure from the assun~ptionof the representative firm typical of general equilibrium models. as in Eliasson (1985) and van Tongeren (1993). the former can gain at the expense of the latter.

111 0. M.-- -- - -.439 REFERENCES Biro Pusat Statistik Indonesia (Indonesian Central Bureau of Statistics).557 0.000 0.125 0.000 0.634 0. Factor Income Distribution Parameters -.185 0. ABBINK. .074 0. LOWN I FARM 2 NFUP 3 NFLO 4 UKUP 5 URLO 6 COMP 7 GOV 8 YI~ Y2h Y3h.434 0. LOWN 1 FARM 2 NFUP 3 NFLO 4 URUP 5 URLO 6 COMP 7 GOV 8 0 0 0 0.665 0. I.138 0.001 0.115 0.341 0.000 0.141 0. COHEN Table A.012 0 0. Jakarta: Biro Pusat Statistik Indonesia. C. 1982. .1 12 0.3.224 0.4 Income Transfer Rates .766 0. I & 11.001 0 0. BRABER AND S. A.096 0. .000 0.187 0.149 0. Social Accounting Matrix lndonesiu 1975.141 0.074 0. Y4 h 0.32 Downloaded By: [Swets Content Distribution] At: 16:31 14 February 2008 G.5 Budget Shares. - - 1 LOWN 2 FARM 3 NFUP 4 NFLO 5 URUP 6 URLO 7 COMP 8 GOV . .098 0..169 0.097 0.100 0.01 1 Table A. Vol.087 0. - - Table A.094 0.

Yogyakarta: Gadjah Mada University Press. a Micrositnulation Approach. THE NETHERLANDS. E.. W. B. The Finn and Firzunciul Markets in the Swedish Micro-to-Macro Model. Journal of Development Econotnics. G e r ~ e r a lEq~rilibriurnModels for Developmerzt Policy..nesia." in H.. 1993. Robinson.D-Thesis. H.. Growth crrzd Incorne Distribution rn Indonesia. Industrialisation arzd trade in In(k.. M. Decaluwe. Rotterdam: Erasmus University. 1982.. 529-568. F...Box 1738. Sunlmer 1988. Eliasson. I.. June 1987. Handbook of Development Ecorzotnics. Erasr7zus University Rotterdurn. Journal ?f'Policy Modelling. 3000 DR.. and Michel. New York: Cambridge University Press. "A Social Accounting Matrix Analysis for the Netherlands". Van Tongcren. J . Chenery and T. de Melo.N. S.. 1990. G. De Econornist. eds. A. homogeneity and dynamics in applied general equilibrium models". P.S. G. Paris: OECD Development Center. and Martens A. Amsterdam: North-Holland. 0. Rottercl~rrn. "Multisectoral Models. PH.. "On closure rules. Kuyvenhoven. S. Dewatripont. Stockholm: Almqvist & Wicksell International. "CGE Modelling and Developing Econonlies: A Concise Empirical Survey of 73 Applications to 26 Countries". S. 1989. Dervis. Poot. Srinivasan. 1990. . 253-272. and Jansen. Corporutes irz nr~Econorny-)vide Model. Adjustment. 65-76. K. Mailirzg Address: Professor S. and Robinson. Thorbecke. Coherz. 1985. I. J.SAM-CGE DEMONSTRATION MODEL Downloaded By: [Swets Content Distribution] At: 16:31 14 February 2008 33 Cohen. Winter 1988.