150Buletin Ekonomi Moneter dan Perbankan, Volume 20, Nomor 2, Oktober 2017

I. INTRODUCTION

One of the most binding constraints facing Indonesia’s economy is limited infrastructure in terms of quantity and quality. Such dire conditions are evidenced by the low infrastructure rating of Indonesia. Data from the Global Competitive Index for 2016-2017, for example, shows that the condition of infrastructure in Indonesia ranks 80th, well behind Malaysia and Thailand, which placed 19th and 49th respectively. The low infrastructure rating of Indonesia has been a major constraint to the realisation of robust and sustainable economic growth as well as economic competitiveness in the country. Furthermore, infrastructure limitations have pushed up the cost of logistics, while undermining economic competitiveness and the investment climate in Indonesia.

The main factor behind weak infrastructure in Indonesia has been limited government spending, particularly since the Asian Financial Crisis in 1997/98. Government spending on infrastructure from 2000-2014 averaged around 2% of GDP, down sharply from 6% of GDP prior to the crisis (Tabor, 2015). In neighbouring peer countries, the ratio of government spending on infrastructure to GDP is roughly three-times that of Indonesia. Moreover, government infrastructure spending in China and India is approximately 10% of GDP.2 Low government infrastructure spending in Indonesia has been attributed to limited sources of financing and the large burden of subsidies, administrative functions and personnel expenditure.

The Government of Indonesia is currently pursuing a program of structural reforms to stimulate economic performance.3 Improving infrastructure through increased spending on investment and less spending on subsidies is part of the structural reform process. Pursuant to the Medium-Term National Development Plan (RPJMN) for 2015-2019, the Government has stipulated the infrastructure targets to be achieved from 2015-2019. A total of 225 infrastructure projects have been upgraded to national strategic projects, covering a broad scope of basic infrastructure, connectivity, electricity as well as communications, water and housing. Furthermore, infrastructure development is not merely restricted to government projects but also involves the private sector through a variety of schemes. Nevertheless, the Government is the dominant investor when striving to achieve the infrastructure development targets set.

Theoretically, larger government investment in infrastructure drives the economy through two channels (IMF, 2014). In the short term, government investment will boost demand through the fiscal multiplier. The impact of government infrastructure investment depends on a range of factors, including economic slack and monetary accommodation. In the medium-long term, however, government investment in infrastructure will affect output from the supply side through increased economic capacity. The demand-side impact of government infrastructure investment depends on the efficiency of that investment.

2http://www.indonesia-investments.com/business/risks/infrastructure/item381

3Structural reforms imply regulatory and institutional improvements to enhance market efficiency (for instance the labor market or goods market), or efforts to stimulate economic growth beyond its potential (Rodrik, 2015). Structural reform policy is very broad and encompasses infrastructure development, tax and subsidy regulations, improving HR quality, wages and employment regulations and bureaucratic reform.

A Dynamic Stochastic General Equilibrium (DSGE) Model to Assess	151
the Impact of Structural Reforms on the Indonesian Economy

From the developments above and considering the important role of infrastructure in terms of nurturing economic growth, quantitative analysis regarding the impact of larger government investment in infrastructure is needed to predict the impact on output and welfare. With a simple approach, the impact of larger government investment in infrastructure on economic growth can be observed using the output multiplier of government spending. Nonetheless, that approach contains several inherent weaknesses, including limited time series data, particularly relating to structural policies. In addition, the simple approach is also lack of the micro-behaviour of individual economic agents and limitations in terms of detecting the linkages between macro variables.

This paper aims to quantitatively analyse the impact of structural reforms – specifically an increase in government spending for investment – on output and welfare in Indonesia using a Dynamic Stochastic General Equilibrium (DSGE) model. Using a DSGE model, the behaviour of economic agents and the linkages between economic players are formulated explicitly. In addition, the DSGE model can not only detect the impact on output, but also on welfare. The DSGE model will be developed consisting of five sectors, namely the household sector, corporate sector, monetary sector, fiscal sector and external sector.

This model adds to the existing DSGE models at Bank Indonesia, which, in general, focus more on banking industry and financial sector modelling, to analyze simulations of monetary policy, macroprudential policy and the monetary and macroprudential policy mix. The model will also complement the Growth Diagnostics model used by Bank Indonesia to analyze the most binding constraints to growth and the impact of structural reforms using the Computable General Equilibrium (CGE)-INDOTERM model (Anugrah et al., 2015). The CGE model used in Growth Diagnostics, however, is unable to capture the dynamics of the monetary sector, fiscal sector and external sector. In addition, the CGE model lacks intertemporal linkages in the correlations between variables.

The remainder of the paper is organized as follows. Section 2 provides a short review on the related literature. Section 3 describes the condition of physical infrastructure in Indonesia. Section 4 explains the structure and specification of the model. Section 5 analyzes the results of the model simulations and Section 6 concludes.

II. LITERATURE REVIEW

This paper relates to previous studies found in the literature. On one hand, this paper is related to studies analyzing the impact of structural reforms on the economies of other countries, primarily using DSGE models. On the other hand, in terms of modelling at Bank Indonesia, this paper is related to several Bank Indonesia studies using DSGE models.

A Dynamic Stochastic General Equilibrium (DSGE) Model to Assess	153
the Impact of Structural Reforms on the Indonesian Economy

Tang at el. (2013) estimate the fiscal multiplier for five ASEAN countries (Indonesia, Malaysia, Philippines, Singapore and Thailand) using a SVAR model. The results show that fiscal policy in the form of lower taxes could have a favorable impact on the five ASEAN members. In Indonesia’s case, the fiscal multiplier of lower taxes is 0.43. In contrast, policy to raise government spending is found to have a negative multiplier effect in Indonesia (-0.34), Singapore (-0.16) and Thailand (-0.27). The main suspect of the negative impact is thought to be imports, which surge due to the small and highly open economies.

Different from the previous two studies on the impact of total government spending on output, this paper examines government spending not as an aggregate but from the perspective of each component, namely spending on investment and spending on consumption. Consequently, the impact of government efforts to increase investment spending can be quantified using the DSGE model developed. In addition, considering the DSGE model is a micro-founded model, the magnitude of impact is not only observed in terms of output but also other macro aggregates (such as inflation and the trade balance) as well as welfare.

2.2. DSGE Models at Bank Indonesia

Like many other central banks, Bank Indonesia has developed several DSGE models. This paper adds to the existing range of DSGE models in Bank Indonesia. After the global financial crisis, DSGE modelling at Bank Indonesia focused on the financial sector. Tjahjono and Waluyo (2010) developed a DSGE model to include the financial accelerator (procyclicality effect), aiming to detect the financial accelerator and its impact on the macroeconomy in the case of a shock, capture the linkages between the monetary and macroprudential policies, simulate the impact of rising fuel prices and analyze the policy mix in the face of economic recession and banking crisis. In the event of a monetary shock, the model with a financial accelerator produces a larger macroeconomic impact than the model without a financial accelerator. An increase in the oil price could potentially lower output, raise inflation and depreciate the rupiah. Furthermore, the impact of monetary policy is more pronounced than that of banking policy in the form of risk-weighted assets (RWA). Nonetheless, the policy mix is still required to maintain the Capital Adequacy Ratio (CAR) as well as prevent and mitigate systemic risk.

Harmanta et al. (2012) developed a DSGE model equipped with the banking sector and applying financial frictions in the household sector in the form of collateral constraints. Their model could be used to assist policy mix formulation as implemented by Bank Indonesia. The simulations show that a hike in the BI Rate would raise bank retail interest rates, reduce lending as well as increase risk-free assets and, therefore, ultimately lower GDP and inflation. An increase in the reserve calculated based on the ratio of changes in national income to changes in government spending or tax revenues.

154Buletin Ekonomi Moneter dan Perbankan, Volume 20, Nomor 2, Oktober 2017

Requirement is shown not to have a significant impact because of excess liquidity in the Indonesian banking sector. An increase in the loan-to-value (LTV) ratio for household/corporate loans, however, could increase output and inflation.

Harmanta et al. (2013) developed further the previous model by adding financial frictions on the entrepreneur side in the form of a financial accelerator, in addition to frictions on the household side. The simulations showed that a hike in the BI Rate would raise the retail interest rate in the banking industry, reduce lending and increase risk-free assets, thereby lowering inflation and GDP. Banks are also exposed to risks as rising non-performing loans (NPL) erode capital. Raising the LTV ratio for households would spur asset purchases, while elevating output and inflation. Raising the Capital Adequacy Ratio (CAR) would lower the loan- to-deposit ratio (LDR), investment and production and, thus, GDP and inflation. The monetary and macroprudential policy mix would create more stable GDP and inflation as well as control consumption and demand for imports.

Harmanta et al. (2014) extended the previous model with an interbank market mechanism as financial frictions on the supply side, while frictions are added to the demand side in the form of collateral constraints and the financial accelerator. The simulations showed that a shock on the interbank market would affect general bank dynamics, particularly bank capital, the Capital Adequacy Ratio (CAR) and loan-to-deposit ratio (LDR). Their model could detect procyclicality and the financial accelerators. Furthermore, the monetary and macroprudential policy mix tended to produce more stable GDP and inflation compared to using just one policy instrument.

III.INFRASTRUCTURE CONDITION IN INDONESIA AND THE GOVERNMENT’S PLAN TO STRENGTHEN INFRASTRUCTURE

3.1. Infrastructure in Indonesia

The condition of infrastructure in Indonesia has improved over the past few years, as reflected by the infrastructure competitiveness index published by the World Economic Forum (WEF). As shown in Figure 1a, the ranking of Indonesia’s infrastructure improved from 92th in 2012 to 60th in 2016. The most salient improvements affected the quality of railways and roads as well as ports (Figure 1b).

156Buletin Ekonomi Moneter dan Perbankan, Volume 20, Nomor 2, Oktober 2017

140
120				Indonesia								Malaysia					Thailand					Vietnam							Philippines
100
80
60
40
20
0

Infrastructure Roads	Railroad	Port	Air transport Electricity	Mobile	Fixed
				telephone	telehphone
				subscriptions	lines

Source : WEF Global Competitiveness Report 2016-2017

Figure 2.

Infrastructure Index of ASEAN Countries

The WEF report also states that the lack of infrastructure in Indonesia was the third biggest constraint to business development over the past few years, after corruption and inefficient government bureaucracy (Figure 3).

Most Problematic factors for doing business

Corruption	11,8
Inefficient government bureaucracy	9,3
Inadequate supply of infrastructure	9,0
Access to financing	8,6
Inflation	7,6
Policy instability	6,5
Poor work ethic in national labor force	6,3
Tax rates	6,1
Inadequately educated workforce	5,6
Tax regulations	4,8
Foreign currency regulations	4,6
Government instability	4,1
Poor public health	4,0
Crime and theft	4,0
Insufficient capacity to innovate	3,7
Restrictive labor regulations	3,7

0

3

6

9

12

Source: WEF Global Competitiveness Report 2016-2017

Figure 3.

Most Problematic Factors for Doing Business in Indonesia

One element of national infrastructure that has received government attention is interregional connectivity. Infrastructure constraints linked to connectivity, such as inadequate ports and poor quality roads, have pushed up the cost of logistics, which is a fundamental problem often found throughout Indonesia. The provision of such infrastructure would lower the cost of transportation and logistics, thereby enhancing product competitiveness and

A Dynamic Stochastic General Equilibrium (DSGE) Model to Assess	157
the Impact of Structural Reforms on the Indonesian Economy

accelerating the pace of the economy. The Logistics Performance Index (LPI), released by the World Bank in 2016, showed that Indonesia placed 63rd out of the 160 countries monitored, with a score of 2.98, down 10 places on the ranking in LPI 2014.

Compared to peer countries in the region, Indonesia remains below Malaysia and Thailand in terms of all LPI aspects, namely customs, infrastructure, international shipping, logistics competencies, tracking and recording as well as timeliness. Indonesia recorded a score of 2.69 for customs, compared to 3.17 in Malaysia and 3.11 in Thailand. For infrastructure, Indonesia received a score of 2.65, while Malaysia and Thailand scored at 3.45 and 3.12, respectively. In terms of international shipping, Indonesia again scored lower than Malaysia and Thailand with scores of 2.90, 3.48 and 3.37 respectively. Concerning logistics competencies, Indonesia received a score of 3.00, compared to 3.34 in Malaysia and 3.14 in Thailand (Figure 4).

LPI 2016 (Rank)

LPI 2016 (Score)

	90	Indonesia		Malaysia	Thailand	Vietnam		Philippines		4,00	Indonesia		Malaysia	Thailand	Vietnam		Philippines
	80									3,80
	70									3,60
										3,40
	60
										3,20
	50
										3,00
	40
										2,80
	30									2,60
	20									2,40
	10									2,20
	0									2,00	LPI (score)	Customs	Infrastructure International		Logistics	Tracking and		Timeliness
		LPI (rank)	Customs	Infrastructure International		Logistics	Tracking and		Timeliness
					Shipments	quality and		tracing						Shipments	quality and		tracing
						competence									competence
	Source: World Bank

Figure 4.

Comparison of Logistics Performance between ASEAN Countries

3.2. Government Investment in Infrastructure

One of the main reasons for the lack of adequate infrastructure in Indonesia is relatively low government investment. The ratio of government investment to GDP in Indonesia from 2006-2014 averaged just 2%, well below Malaysia and Thailand at 9.9% and 6.4% respectively (Figure 5). In 2015, the government increased investment significantly from 2.41% to 3.23% of GDP, thereby surpassing government investment in the Philippines at 2.81% but still trailing Malaysia and Thailand.

158Buletin Ekonomi Moneter dan Perbankan, Volume 20, Nomor 2, Oktober 2017

%GDP

12
10
8
6
4
2
0	Indonesia		Malaysia			Thailand		Philippines
2006	2007	2008	2009	2010	2011	2012	2013	2014	2015

Source : World Bank

Figure 5. Comparison of Government

Investment between ASEAN Countries

In order to achieve higher and sustainable economic growth, the government is committed to improving the quality and quantity of infrastructure. Such commitment is evidenced by the infrastructure development targets to be achieved by 2019, pursuant to the Medium-Term National Development Plan (RPJMN) for 2015-2019 (Table 1).

Table 1.

Infrastructure Development Targets in RPJMN (2015-2019)

Indicator	2014 Baseline	2019 Target
Basic infrastructure of energy and electricity
Electrification ratio (%)
	84.1	96.6
Electricity consumption per capita (kWh)	843	1200
Establishment of Floating Storage Regasification Unit (unit)	2	3
Gas pipelines (km)	11,960	17,690
Establishment of gas station for transportation (unit)	40	118
Use of natural gas (household connection)	102 Thousand	1 Million
Addition of petroleum refineries (unit)	-	2
Basic Infrastructure of residential areas
Clean water access
	68.5%	100%
Adequate sanitation access	60.5%	100%
Urban slums neighborhood	37.407 Ha	0 Ha
Housing backlog	13.5 Million	6.8 Million
Connectivity
Well-maintained National Roads	94%	100%
Logistic cost (%GDP)	23.5%	19.2%
Urban public transport market share	23%	32%

A Dynamic Stochastic General Equilibrium (DSGE) Model to Assess 159 the Impact of Structural Reforms on the Indonesian Economy

Table 1.

Infrastructure Development Targets in RPJMN (2015-2019)

Indicator	2014 Baseline		2019 Target
Boardband coverage to regencies/ cities	72%		100%
Water Durability
Standard water infrastructure capacity	51.4	M3/Sec	118.6M/Sec
Water storage capacity per capita	62.3	M3/Capita	78.36	M3/Capita
Dam irrigation coverage	11%		20%
Surface Irrigation Networks	7,145 Million Ha		7,914	Million Ha
Flood control capacity	5-25 Years		10-100 Years
Source: RPJMN 2015-2019

To achieve those targets, the government has already stipulated the infrastructure to be built during from 2015-2019. Accordingly, a total of 225 infrastructure projects have been designated as national strategic government projects. In addition, the government is also developing a 35,000MW national electricity project that is separate from the other national strategic projects. Of the total, around 46 projects are located in Sumatra, 89 in Java, 24 in Kalimantan as well as 16 in Bali and Nusa Tenggara. There are a further 28 projects located in Sulawesi, 13 in Maluku and Papua and 10 more distributed in various provinces. The national strategic projects are based on Presidential Decree No. 3/2016 concerning the Acceleration of National Strategic Project Implementation, as well as Presidential Instruction No. 1/2016 regarding the Acceleration of National Strategic Project Implementation.

The acceleration of national strategic projects is pursuant to Presidential Decree No. 3/2016, which regulates licensing and non-licensing, government procurement of goods/ services, discretional debottlenecking, spatial planning, land acquisition, government funding and guarantees, as well as the use of domestic apparatus. Data from the Coordinating Ministry of Economic Affairs showed that in June 2016, a total of 86 projects (44%) had completed ground breaking and entered the construction phase, while the remaining 139 projects were still at the planning phase or set to commence implementation. Furthermore, per June 2016, of the total planned 35,000MW electricity project, as much as 17,800MW had been signed to project contracts.

To expedite infrastructure development, the government introduced fiscal, institutional and regulatory reforms. The institutional reforms included strengthening The Committee for Acceleration of Priority Infrastructure Delivery (KPPIP) and the Presidential Decree on National Strategic Projects. In addition, the government also reclassified the projects to ensure a greater focus on project completion. Of the 225 existing national strategic projects and one electricity program, pursuant to Coordinating Minister of Economic Affairs Decree No. 12/2015, the government has designated 30 priority projects to receive special attention from KPPIP.

160Buletin Ekonomi Moneter dan Perbankan, Volume 20, Nomor 2, Oktober 2017

Infrastructure investment, which has become a government priority, is still constrained by a lack of funding. Data from the National Development Planning Agency, as contained in the MediumTerm National Development Plan (RPJMN) for 2015-2019, stipulates the funding requirement for the priority infrastructure projects from 2015-2019 at Rp5,519.4 trillion. The funding requirement consists of government ministry/agency spending and regional transfers (national budget and local government budgets) amounted to Rp2,760.9 trillion or around 50.02%, state-owned enterprises amounted to Rp1,066.2 trillion or 19.32%, and private participation amounted to Rp1,692.3 trillion or 30.66% (Table 2).

Table 2.

Funding Requirement for Infrastructure in RPJMN 2015-2019

	State	Regional
Sector	Budget	Budget	SoE (BUMN)2	Private3	Total
	(APBN)1	(APBD)
Road	340.0	200.0	65.0	200.0	805.0
Railway	150.0	-	11.0	122.0	283.0
Sea transportation4	498.0	-	238.2	163.8	900.0
Air transportation	85.0	5.0	50.0	25.0	165.0
Land transportation	50.0	-	10.0	-	60.0
City transportation 5	90.0	15.0	5.0	5.0	115.0
Electricity 6	100.0	-	445.0	435.0	980.0
Energy (oil and gas)	3.6	-	151.5	351.5	506.6
Information and communication	12.5	15.3	27.0	223.0	277.8
technology
Water resources	275.5	68.0	7.0	50.0	400.5
Drinking water and sanitation	227.0	198.0	44.0	30.0	499.0
Housing	384.0	44.0	12.5	87.0	527.5
Total Infrastructure	2,215.6	545.3	1,066.2	1,692.3	5,519.4
Percentage	40.14%	9.88%	19.32%	30.66%	100.00%

Source: Ministry of National Development Planning Notes:

1)Expected state budget support

2)Expected state-owned enterprise support

3)Maximum capacity of private finance through Public-Private Partnership including business to business scheme

4)Increase due to the addition of sea toll component and routine cost

5)Allocated for rail and road-based urban transport

6)PT PLN has the capability to finance Rp 250 T, the remaining cost requires state investment

A Dynamic Stochastic General Equilibrium (DSGE) Model to Assess	161
the Impact of Structural Reforms on the Indonesian Economy

In line with government efforts to improve the quality and quantity of infrastructure, the national budget allocation to infrastructure has increased significantly. Based on data from the Ministry of Finance, government spending on infrastructure in 2015 reached Rp290.3 trillion (3.23% of GDP), up 40.5% from the previous year. A further increase was observed in 2016, with the budget allocation went up to Rp313.5 trillion. In 2017, Rp346.6 trillion has been earmarked for infrastructure. With the growing infrastructure budget through to 2019, government spending on infrastructure is expected to reach 4.21% of GDP (Figure 6). As an aggregate for 2015-2019, the government’s infrastructure budget is predicted to reach Rp1,796.2 trillion.

	Rp Trillion												% GDP
500													4,21%	4,50%
		Infrastructure Spending (Rp level)
450													3,67%	4,00%
		Infrastructure Spending (% GDP)
400												3,33%	3,50%	3,50%
											3,23%
350
														3,00%
300										2,41%
														2,50%
250									1,90%
		1,93%	1,81%	1,87%	1,89%	1,79%	1,66%	1,70%						2,00%
200
														1,50%
150	1,18%
100														1,00%
50														0,50%
0														0,00%
	2005	2006	2007	2008	2009	2010	2011	2012	2013	2014	2015	2016*	2017** 2018** 2019**

Source: Ministry of National Development Planning, Ministry of Finance, BI

Figure 6.

Goverment Spending on Infrastructure 2015-2019

IV. DSGE MODEL

The model developed in this paper is based on Ganelli and Tervala (2015), where government consumption and investment are inputted separately into a New Keynesian DSGE Model. The main modification to the model is to assume a small open economy in order to enrich the impact assessment of structural reforms in Indonesia. In general, the model also contains the features of the standard DSGE model based on the model of Gali and Monacelli (2005). The structure of the model is presented in Figure 7.

162Buletin Ekonomi Moneter dan Perbankan, Volume 20, Nomor 2, Oktober 2017

Government

Paying taxes

Producing goods for

government

Providing public consumption

Providing infrastructure

Supplying labor and consuming goods

Household

Firms

Export

Producing consumption goods

Import

Central Bank

World

Conducting monetary policy

Figure 7.

Structure of the Model

4.1. Households

4.1.1. Household Utility Maximization

All households are assumed to have the same preferences. Households maximize their utility function based on their choice of consumption level, 𝐶𝑡, and time spent resting (outside working hours, 𝑁𝑡), given the public goods , 𝐺𝑡𝐶, provided by the government.

1

where β is discount factor, σ is intertemporal elasticity of substitution of household, φ is labor supply elasticity, and ν is weight of public consumption.

Households receive an income from wages for providing labor to a firm, 𝑊𝑡𝑁𝑡, income from term deposits, (1 + 𝑖𝑡)𝐷𝑡, and dividends from their own businesses, Π𝑡. The income is used to pay taxes, 𝑇𝑡, and spend on consumption. Therefore, the budget constraints faced by households is the following:

𝑃𝑡𝐶𝑡 + 𝑄𝑡+1𝐷𝑡+1 = 𝐷𝑡 + 𝑊𝑡𝑁𝑡 + Π𝑡 − 𝑃𝑡𝑇𝑡

2

where is a stochastic discount factor. In the budget constraints, spending on consumption and taxes are multiplied by prices to form the nominal value.

A Dynamic Stochastic General Equilibrium (DSGE) Model to Assess	163
the Impact of Structural Reforms on the Indonesian Economy

Optimizing the objective function (1) with the budget constraint (2) produces the optimal solution (First Order Condition) for consumption, 𝐶𝑡, and working time, 𝑁𝑡, as follows:

3

4

Households determine their level of consumption at time t, 𝐶𝑡, by considering the expected spending on consumption in future periods, , price changes, , and the discount factor.

Meanwhile, households choose to work, 𝑁𝑡𝜑 , in line with the wages received, 𝑊𝑡, to meet the consumption requirement.

4.1.2 Optimum Consumption Allocation of Households

Household consumption, 𝐶𝑡, is a composite index of domestic goods, 𝐶𝐻,𝑡, and imported goods 𝐶𝐹,𝑡.

where 𝛼 is the consumption home bias parameter and 𝜂 is elasticity of substitution between domestic and imported goods. The respective consumption of domestic goods, 𝐶𝐻,𝑡, and imported goods, 𝐶𝐹,𝑡, constitute aggregate consumption of various types of goods 𝑗, namely:

where 𝜀 and 𝛾 are the elasticity of substitution between varieties of domestic and foreign goods respectively.

Households maximize total consumption by determining the consumption allocation for various types of goods, 𝑗, for each level of spending, 𝑍𝑡.

164Buletin Ekonomi Moneter dan Perbankan, Volume 20, Nomor 2, Oktober 2017

𝑥 = 𝐻, 𝐹, and 𝜗 = 𝜀, 𝛾. The optimal solution (FOC) for this problem is for households to allocate the consumption of goods, 𝑗, according to relative prices of other goods and the substitution elasticity between goods, which can be expressed as follows:

Furthermore, households will also maximize their consumption allocation between domestic and international goods. With the same settlement measures as before, the optimal solution in terms of consuming domestic and international goods will be obtained as follows:

where

5

is the consumption price index.

4.1.3.Relationship between Domestic and International Prices and Consumption

A price comparison between domestic and international goods (terms of trade) can be expressed

as, or in log-linear form as 𝑠𝑡 = 𝑝𝐹,𝑡 − 𝑝𝐻,𝑡. Meanwhile, the log-linear form of the consumer

price index equation (5) in a steady state condition produced 𝑝𝑡 = (1 − 𝛼)𝑝𝐻,𝑡 + 𝛼𝑝𝐹,𝑡, therefore 𝑝𝑡 = 𝑝𝐻,𝑡 + 𝛼𝑠𝑡. Thus, a change in the Consumer Price Index (CPI) is determined by a change of domestic prices and a change in the terms of trade (ToT).

𝜋𝑡 = 𝜋𝐻,𝑡 − 𝛼(𝑠𝑡 − 𝑠𝑡− 1)

International commodity prices in the domestic currency (𝑝𝐹,𝑡) are the same as global prices in a foreign currency (pt*) multiplied by the nominal exchange rate pF,t = et + pt*. Therefore, the terms of trade can also be formulated as follows:

A Dynamic Stochastic General Equilibrium (DSGE) Model to Assess	165
the Impact of Structural Reforms on the Indonesian Economy

Meanwhile, the real exchange rate can be defined as a change in international prices in the domestic currency divided by domestic prices (Consumer Price Index):

𝑞𝑡 = 𝑒𝑡 + 𝑝𝑡∗ − 𝑝𝑡

Thus, the relationship between the real exchange rate and the terms of trade can be defined as follows:

𝑞𝑡 = (1 − 𝛼)𝑠𝑡

Furthermore, assuming linearity, changes in the real exchange rate and global growth to meet domestic demand will determine the magnitude of change in domestic consumption:

4.1.4. Market Clearing between Consumption and Production

In equilibrium, companies produce goods at the same amount with government spending and household consumption such that:

Using the terms of trade and exchange rate equations, the equation above can be expressed as:

4.2. Firms

4.2.1. Minimizing Production Costs

Firms are assumed to have factors of production consisting of labor and public infrastructure provided by the Government. A production function is also assumed to be linear with respect to technological progress. Each firm produces various different products 𝑗:

𝑌𝑡(𝑗)

𝐴𝑡(𝐾𝑡𝐺)Φ𝑁𝑡(𝑗)

6

where Φ is output elasticity to public infrastructure. Parameter Φ is assumed to have a positive value, implying that the production function has an increasing returns with respect to public infrastructure. The problem faced by firms in finding the real marginal cost is to minimize the real total cost given certain level of production. Firms choose number of labors to minimize the real total cost as follows:

166Buletin Ekonomi Moneter dan Perbankan, Volume 20, Nomor 2, Oktober 2017

7

Optimization of the objective function (7) with the constraint (6) produces the optimum function of the marginal cost 𝑀𝐶𝑡 as follows:

8

Therefore, the marginal cost is found to be equivalent to spending on real wages divided by the marginal product of labor 𝐴𝑡(𝐾𝑡𝐺)Φ.

4.2.2. Maximizing Profit and Price Setting

In this model, each firm is assumed to produce various different products 𝑗, thus giving the company bargaining power to change the prices of its products 𝑃𝑡(𝑗). A portion of firms are assumed to set prices flexibly with a price index 𝑃𝑡, and some firms strive to seek optimal prices Pt*. With the presence of price rigidity, the process of setting optimal prices follows the framework introduced by Calvo (1983).

9

10

The optimal solution of the objective function (9) with the constraint of equation (10) produces the following equation:

11

The value of is called the price mark-up and determined by the elasticity of substitution

between domestic products 𝜀.

Each company can optimize or adjust prices with a probability of 1 − 𝜃 (independent) in each period, thus 𝜃 < 1 can be interpreted as the price rigidity index. Therefore, the aggregate price 𝑃𝑡 is as follows:

A Dynamic Stochastic General Equilibrium (DSGE) Model to Assess	167
the Impact of Structural Reforms on the Indonesian Economy

By denoting that inflation, then equation (11) can be expressed in the final log-linear form:

12

Equation (12) shows that a change in the optimal price at time 𝑡 (𝜋𝑡) is the weighted average of the marginal cost in the current period and future price changes. In addition, equation (8) also shows that an increase in the stock of public infrastructure would lower the optimal price.

4.3.Fiscal and Monetary Policy

4.3.1.Fiscal Policy

The government receives revenue from taxes, which is then used for consumption and investment.

𝑃𝑡𝑇𝑡

𝐺𝑡𝐼 + 𝐺𝑡𝐶

For simplicity, taxes, 𝑇𝑡, are non-distortionary. If the government expenditure financed by taxes is distortionary, the impact of expansionary fiscal policy would be less pronounced and, thus, affect the impact on welfare.

Government spending on consumption and investment is modelled using the AR(1) dynamics:

𝑔̂𝑡𝐶 𝜌𝐶𝑔̂𝑡𝐶+ 𝜀𝑡𝐶 𝑔̂𝑡𝐼 𝜌𝐼𝑔̂𝑡𝐼+ 𝜀𝑡𝐼

Parameters 𝜌𝐶 and 𝜌𝐼 represent the persistence of government consumption and investment shocks. Meanwhile, 𝜀𝑡𝐶 and 𝜀𝑡𝐼 represent i.i.d shocks on government spending with a normal distribution and standard deviation 𝜎𝑔.

Like Ganelli and Tervala (2015), private capital is omitted from the model, thus providing greater focus on the impact of government investment spending. The stock of public infrastructure is modelled using the following equation:

The stock of public infrastructure depreciates at 𝜆 and expands in line with government investment spending.

A Dynamic Stochastic General Equilibrium (DSGE) Model to Assess	169
the Impact of Structural Reforms on the Indonesian Economy

In addition, the output multiplier can also be measured based on the net present value of fiscal multiplier (NPVFM), which is calculated as a ratio of the net present value of the change in output to the net present value of the change in government spending for a given period.

where 𝑥 = 𝐶, 𝐼. In this paper, we set number of periods at 2,000 quarters, following Ganelli and Tervala (2015).

To measure the impact of government spending on welfare, the net present value of the impact of fiscal expansion (𝜏𝑡) on welfare is measured as the part of consumption households are willing to release due to fiscal expansion. The value of 𝜏𝑡 is calculated as follows. Based on the household utility function, the net present value of household welfare without fiscal expansion is as follows:

16

The net present value of household welfare without fiscal expansion is as follows:

17

From equations (3.16) and (3.17), the net present value of the impact of fiscal expansion on welfare can be solved as follows:

Therefore, the welfare multiplier impact is obtained based on the cumulative change in utility (multiplied by the discount factor) and divided by the cumulative government spending.

170Buletin Ekonomi Moneter dan Perbankan, Volume 20, Nomor 2, Oktober 2017

4.6.Optimal Solution of the Model

Optimal solution to the model, based on household, corporate, central bank, government and external conditions, as described above, can be summarised in Table 3:

	Table 3.
	Optimal Solution and Log-linearisation
No	Optimal Solution (First Order Condition)	Log-linear

1 Euler Equation:

2 Philips Curve:

3	Marginal cost:
4	Consumption:
5	Inflation:			𝜋𝑡 = 𝜋𝐻,𝑡 + 𝛼(𝑠𝑡 − 𝑠𝑡− 1)
	𝜋𝑡 = 𝜋𝐻,𝑡 + 𝛼(𝑠𝑡 −			𝑠𝑡− 1)
6	Terms of trade:
	𝑠𝑡 = 𝑠𝑡−1 + (𝑒𝑡 −			𝑒𝑡− 1) + 𝜋𝑡∗ + 𝜋𝐻,𝑡
7	Taylor rule:
8	Market Clearing:

9	Capital Law of Motion:
	𝐾𝑡𝐺+1	= (1 −	𝜆)𝐾𝑡𝐺 + 𝐺𝑡𝐼
10	Employment:			𝑦̂𝑡= 𝑎̂𝑡+	Φ𝑘̂𝑡𝑔+ 𝑛̂𝑡
	𝑌𝑡(𝑗)	𝐴𝑡(𝐾𝑡𝐺)Φ𝑁𝑡(𝑗)
11	Trade Balance:
12	Inflation:			𝜋𝑡 = 𝑝𝑡 −	𝑝𝑡−1
	𝜋𝑡 = 𝑝𝑡 − 𝑝𝑡−		1	𝜋𝐻,𝑡 𝑝𝐻,𝑡 − 𝑝𝐻,𝑡−1 𝜋𝑡∗ 0
	𝜋𝐻,𝑡	𝑝𝐻,𝑡 −	𝑝𝐻,𝑡−1 𝜋𝑡∗	0

A Dynamic Stochastic General Equilibrium (DSGE) Model to Assess 171 the Impact of Structural Reforms on the Indonesian Economy

		Table 3.
	Optimal Solution and Log-linearisation
No	Optimal Solution (First Order Condition)		Log-linear
13	Technology Shock:		𝑎𝑡 = 𝜌𝑎𝑎𝑡−1 + 𝜀𝑡𝑎
	𝑎𝑡 = 𝜌𝑎𝑎𝑡−1 + 𝜀𝑡𝑎
14	Government Investment Shock:
15	Government Investment Shock:

4.7. Steady State Value

In this paper, the steady state of working hours 𝑁̅ is assumed at 1/3, meaning that the labor force spends one-third of its time at work. This value is consistent with Colley (1995). Meanwhile, the steady state of public capital stock ̅𝐾̅𝐺 is normalised to 1, thus, based on the production function (6), the steady state of output is given by 𝑌̅= 𝑁̅. Furthermore, based on the market

clearing equation, the steady state of consumption is . Based on the capital depreciation equation, the steady state value of government intervention is the same as the rate of depreciation ̅𝐺̅𝑡̅𝐼= 𝜆.

4.8. Parameterization

The parameter values used in the model are determined through a calibration and estimation process. The discount factor 𝛽 is set at 0.99. The model is interpreted as a quarterly model, and therefore the assumption of 𝛽 = 0.99 is equivalent to an annualized interest rate of 4%. The elasticity of substitution between domestic and international goods (𝜂) is assumed at 3.5, which is consistent with the findings of Obstfeld and Rogoff (1998). Meanwhile, the elasticity of substitution between international goods 𝛾 is assumed equals to 1, in line with Gali and Monacelli (2005). The elasticity of substitution of labor supply as 𝜑 is assumed at 1.97, the same as Sin (2016) and not far different from Ganelli and Tervala (2015). The wealth effect of labor supply 𝜎 is assumed equals to 2.

The value of the Calvo parameter 𝜃 = 0.5 is assumed at 0.5, which is the value often used in numerous New Keynesian models and the same as Tjahjono et al. (2009) for the Bank Indonesia Structural Macromodel (BISMA). This assumption leads to a lag of the prices adjustment in two periods (six months). The value of the home bias parameter (1 − 𝛼) is obtained using the average ratio of imports to GDP from 2005-2015, namely 𝛼 = 0.2. Interest rate smoothing 𝜌𝑖 is assumed to have a value of 0.75, while the respective weights of inflation and output are 𝜙𝜋

=1.9 and 𝜙𝑦 = 0.25. The Taylor rule parameters are the same as that estimated by Harmanta et al. (2012).

172Buletin Ekonomi Moneter dan Perbankan, Volume 20, Nomor 2, Oktober 2017

Fiscal policy is assumed to have an investment and consumption shock persistence of 𝜌𝐼 = 𝜌𝐶 = 0.75, respectively, mirroring the findings of Iwata (2013). Meanwhile, the share of government consumption relative to domestic consumption 𝜈 is assumed at 0.15, in line with the average value for Indonesia during the period from 2011-2015. Public infrastructure is assumed to depreciate each quarter at the rate of 𝜆 = 0.025, which is the same rate of depreciation as that used in the model of Bom and Ligthart (2014). The output elasticity of public infrastructure stock, 𝜙𝑘, is a key parameter in this paper, with 𝜙𝑘 = 0.24, in line the findings of Aschauer (2000) for the case of low and middle income countries. A summary of the model parameters is presented in Table 4.

		Table 4.
		Calibrated Value of Parameters
Parameter	Baseline value	Description
𝛽	0.99	Discount factor
𝜂	3.5	Elasticity of substitution between domestic and imported goods
𝛾	1	Elasticity of substitution between varieties of foreign goods
𝜑	1.97	Frisch elasticity of labor supply
𝜎	2	Intertemporal elasticity of substitution of household
𝜃	0.5	Calvo Parameter
𝛼	0.2	Home bias parameter
𝜌𝑖	0.75	Interest rate smoothing
𝜙𝜋	1.9	Taylor rule coefficient for inflation
𝜙𝑦	0.25	Taylor rule coefficient for output
𝜌𝐼	0.75	Persistency of public investment shock
𝜌𝑐	0.75	Persistency of public consumption shock
𝜈	0.15	Weight of public consumption
𝜆	0.025	Depreciation rate of public infrastructure
𝜙𝑘	0.24	Output elasticity of public infrastructure

The results of parameter calibration are used as the priors for the estimations using Bayesian approach, as explained in Quintana (2012). In this paper, the estimated parameters are 𝜎, 𝜑, 𝛾, 𝜃 and 𝜂. More details, the prior distribution, distribution type and posterior distribution of the parameter estimations are presented in Table 5.

A Dynamic Stochastic General Equilibrium (DSGE) Model to Assess 173 the Impact of Structural Reforms on the Indonesian Economy

Table 5.

Estimated Value of Parameters

Parameter	Description	Distribution	Prior distribution				Posterior
							distribution
			Mean		Std. Dev.
𝜃	Calvo parameter	Beta		0.67		0.1	0.51
𝜎	Intertemporal elasticity of	Gamma		2.00		0.1	2.70
	substitution of household
𝜑	Frisch elasticity of labor supply	Gamma		1.97		0.1	2.67
𝛾	Elasticity of substitution between	Gamma		1.00		0.2	2.24
	varieties of foreign goods
𝜂	Elasticity of substitution between	Gamma		3.50		0.8	1.75
	domestic and imported goods

V. Simulation Results

This section describes the dynamics of the impulse functions. The discussion focuses on the simulations of fiscal policy, that is, government consumption and government investment shocks. In line with the design of the model, in addition to the impact of fiscal policy on macroeconomic aggregates, the impact on household welfare is also measured. Figure 8 shows the simulation results of fiscal expansion, where the solid line indicates the impact of the government consumption shock and the dashed line illustrates the impact of the government investment shock. In general, the shock produces deviation from the steady state of each respective variable, excluding the trade balance, which is represented by deviation from the steady state of output, and inflation is in percentage change from the previous year.

5.1. Government Consumption

In the short run, a 1% increase in government consumption would potentially results in a 0.04% higher economic growth above the baseline without policy (Figure 9), driven by an increase in aggregate demand, the wealth effect and working hours. The government would raise taxes to offset the increase in consumption, while the labor force would be offset by an increase in working hours to earn higher wages. Although households reduce their consumption, the aggregate impact of an increase in government consumption remained positive. A cumulative increase in government consumption would raise economic growth by 0.03% (Table 6).

The trade surplus would initially declines, prompted by an increase in imports to meet government consumption. Thereafter, however, the decrease in household consumption due to raised taxes and an improvement in the terms of trade in the medium-long term, would improve the trade balance. The hump-shaped distribution of the trade balance is in line with Iwata’s (2013) study.