It is useful to provide a brief overview of the development stages of KGEMM, as these stages shaped the structure of the current version of the model. As mentioned previously, KGEMM has been developed to have a better representation of the Saudi Arabian economic (sectoral and macro) and energy relationships. The main motivation for developing it was that there was no available model (including subscription based) that properly represented the Saudi Arabian economy and could comprehensively inform the policy decision-making process.

The model presented in this book is the fifth version of KGEMM. The first version of the model was built in early 2014 by Frederick L. Joutz, Fakhri J. Hasanov, and John Qualls, researchers of the KGEMM team at KAPSARC. It was built upon the Saudi Arabian module of the Oxford Economics Global Economic Model (OEGEM). KGEMM enhanced the OEGEM’s Saudi Arabian module by addressing its key limitations, including its oversimplified representation of the Saudi economy, which would prevent it from being used to comprehensively inform the policymaking process.Footnote 1 KGEMM differs from the Saudi Arabian module of OEGEM considerably.Footnote 2 The main feature of the second version of KGEMM was the development of a detailed energy block representing 14 energy demand relationships by energy type and customer. The third version of KGEMM had estimated production function relationships for economic activity sectors, and thus derived sectoral output gaps feeding into consumer price index (CPI) equations for 12 household consumption basket items. It also had employment demand equations for economic activity sectors estimated as a function of output and wage (see Hasanov et al. 2021). Finally, it had more detailed external sector representations. For example, it linked the export of Saudi Arabian refinery oil to 45 individual countries’ refinery oil demand, which offers the opportunity to simulate the impact of the global demand for oil and thus, environmental implementations and energy transitions on the Saudi economy. The third version of KGEMM was published as Hasanov (2020). In the fourth version of KGEMM, sectoral wage relationships were econometrically estimated as a function of the labor productivity and output price mainly. Additionally, sectoral investment relationships were developed using econometrically estimated investment demand equations, as a function of output, interest rate, and exchange rate (see Javid et al. 2022). That version also considered substitution effects in energy demand equations. Finally, the fifth version of KGEMM has a number of developments that considerably differentiates it from the previous versions. These developments include but are not limited to the following: CO2 block, which assesses carbon emissions of the energy block, that is, each of 15 energy products; more detailed external sector represented by enhanced non-oil export equation (Hasanov et al. 2022c), developed oil refinery export equation, and outflow remittances equation (Javid and Hasanov 2022); representations of the supply-side of energy, that is, fuel mix components for electricity generation as well as renewable energy represented by solar energy (Elshurafa et al. 2022; Hasanov et al. 2022a); representations of the petrochemical sector through mainly estimated output, investment, employment, energy demand, and feedstock (ethane, methane, naphtha, liquified petroleum gases) demand equations; imports of goods disaggregated into capital, intermediate, and consumer goods, each estimated econometrically as a function of domestic demand and real exchange rate. Details of the developments in the fifth version of KGEMM can be noticed in the description of each block below in this section as well as in behavioral equations and identities given in the following section. Figure 6.1 illustrates the structure of the fifth version of KGEMM.

Fig. 6.1
A network diagram represents price, monetary, fiscal, external, population and age cohort, carbon dioxide, energy, wage and labor, and real in the center with maximum connection.

KGEMM structure

Full size image

The model has nine blocks interacting with each other to represent the Saudi Arabian energy–economic–environment relationships. What follows is a brief block-by-block description of KGEMM’s structure.

6.1 Real Block

This block can be broadly divided into demand-side and supply-side representations.

6.1.1 Real Block: Demand Side

Conventional MEMs, many of them for the Saudi Arabian economy, treat demand on an aggregate level. However, as mentioned earlier, aggregate demand in KGEMM is broken down into intermediate demand, final demand, and total demand for 13 economic activity sectors. The intermediate demand is modeled as the demand of all these economic activities for each other’s goods and services using coefficients derived from the input–output table for Saudi Arabia. Also, the final demand components of gross domestic product (GDP), such as private consumption, government consumption, investment, and exports are disaggregated into the economic activities using the coefficients derived from the input–output table. In addition, the investments are further broken into government, oil sector, and non-oil sector investments. While exports are represented by non-oil, oil, and service exports. The total demand for a given economic activity sector is the sum of its intermediate demand and final demand. Such a detailed framework makes the model able to distinguish intermediate, final, and total demand effects in different sectors of the economy. For example, the effects of government investment in the construction and transportation sectors are not identical, but sector-specific, each represented by its own coefficient. This allows a modeler to quantify sector-specific effects of the government policies.

Private consumption is econometrically estimated as a function of private disposable income, interest rate, and wealth using cointegration and ECM methods by Hasanov et al. (2022b) and incorporated into KGEMM. Note that the private disposable income data is not available from official agencies and hence has been constructed by the authors using the System of National Accounts framework.

Investment is the sum of oil and non-oil private and government investments as mentioned above. Non-oil private investment is the sum of domestic and foreign private investments. The latter one is the sum of foreign direct, foreign portfolio, and foreign other investments—all coming from the external block. The domestic private investments are the sum of private investment in eight non-oil economic activity sectors. Sectoral private investments are econometrically modeled as a function of sectoral output, interest rate, and exchange rate by Javid et al. (2022) over the period 1989–2017. Later, the KGEMM team updated estimations till 2019 and expanded the estimation coverage by modeling petrochemical sector investment.

The other final demand components, i.e., government consumption, exports and imports, will be discussed in their respective blocks later. These indicators link the real block to the fiscal and external blocks.

The economic activities are econometrically estimated, using the demand-side approach, where demand for a given economic activity sector is a function of total demand for this sector and the sector’s demand for energy. This approach comes from the input–output framework and additionally includes energy demand. This is very similar to the modeling approach taken by Bradley et al. (1995) for the European countries. The purpose of including energy demand variable in the estimations is to measure the explicit effects of energy on economic activities. This makes it possible to use the model to analyze the impact of the domestic energy sector-related reforms on various economic activities. These estimations link the real sector to the energy and price blocks.

6.1.2 Real Block: Supply Side

The supply side of the real block mainly contains production functions for the economic activities, which estimate the potential outputs of the activities as functions of capital stocks and employment mainly, alongside technological change proxied by the time trend and other explanatory variables. This is consistent with the theory of production (Cobb and Douglas 1928; Douglas 1976). We use the Cobb–Douglas type as the form of the production function because it has been recommended and widely used in macroeconometric modeling (e.g., see Welfe 2011).

The capital stock for a given economic activity sector has been constructed using investment in the sector, the sector-specific depreciation rate, and assumed initial capital stock in the perpetual inventory method framework (Collins et al. 1996; Nehru and Dhareshwar 1993; Hall and Jones 1999; Arezki and Cherif 2010). The relationships representing employment for economic activity sectors, used in the production function estimations, will be discussed in the wage and labor block.

The output gaps of the economic activities have been calculated using the identities, which express the differences between the actual outputs and potential outputs coming from the estimated production functions. The gaps feed into the behavioral equations for the CPIs of the different household consumption basket items, as discussed in the price block.

Demand-side and supply-side breakdowns of the entire economy into economic activity sectors can provide useful information about sectoral compositions and changes in the structure of the economy. This is very important for diversification and local content purposes, key targets of SV2030, as articulated in the National Transformation Program (SV2030 2019b).

Section 7.1 provides additionally definitional relationships for disposable income and wealth of households, total final expenditure, and domestic demand. The sub-section also presents the nominal value added of the economic activity sectors and the aggregation of the economic activity sectors into large sectors, such as the service sector, the industry sector, the oil sector, and the non-oil sector.

6.2 Fiscal Block

Total government expenditure is the sum of government capital and current expenditure in KGEMM. The former is a function of the one-period lagged capital expenditure, the relative increase in government investment, and the remainder of capital expenditure. The government current expenditure is the sum of the government’s five current spending items, namely, wages, salaries, and allowances; administrative expenses; maintenance and operational costs; transfers to the private sector; and other current expenses. Government consumption is the sum of the first three current spending items above. Each of these five current spending items alongside government investment spending is econometrically estimated using government revenues as an explanatory variable.

Total government revenue is the sum of the government’s oil and non-oil revenues. The former is linked to oil exports from the external block. The government’s non-oil revenues are the sum of revenues from energy sales, value added tax (VAT), the expatriate levy, the Umrah and Hajj visa fees, other visa fees, tax on international trade and transactions (which is effectively customs duty fees on imports as Saudi Arabia does not apply tax or fee on exports), taxes on income, profits, and capital gains, and other non-oil revenues. The latter is the balancing item and includes collections such as traffic fines, the idle land tax, the luxury good tax, government investment returns, other non-oil sector related fees, tariffs, and collections. Energy sales are linked to the economy’s total energy consumption in monetary measure coming from the energy block. VAT revenues is formed by the final consumption, VAT rate, and VAT collection efficiency ratio. The last two are policy variables for the fiscal authority. The Umrah and Hajj visa fees and other visa fees are linked to developments in the exports travel service, while the expatriate levy is tied to developments in the migrant population. The taxes on income, profits, and capital gains are linked to the changes in the non-oil sector’s activity. The total government revenue in the model has been structured according to the stylized facts of the Saudi Arabian economy. This allows us to model the impacts of oil market changes, which are mostly exogenous, as well as the impacts of internal policy decisions for the energy prices reform, VAT, expatriate fees, and other revenue components—the key initiatives of the Fiscal Balance Program (FBP) of SV2030 (2019a)—on the Kingdom’s economy.

Lastly, the block contains the government’s budget balance: the difference between its total revenues and total expenditures. It also contains the government’s non-oil budget balance: the difference between the government’s non-oil revenues and its total expenditure. The latter is another stylized fact of the economy, and its purpose in the model is to show to what extent non-oil sector revenues can finance government spending. The government can then take measures to increase the efficiency of both its spending and revenue collection and fiscal consolidation—the key objectives of the FBP (SV2030 2019a). The block also has general government debt, which is econometrically modeled as being dependent on government balance. General government gross debt is the sum of its past value and general government debt. Section 7.2 expresses the relationship described above.

6.3 Monetary Block

The monetary block of the model is fairly small, mainly because of the nature of monetary policy in Saudi Arabia, another stylized fact of the economy. The exchange rate of the Saudi riyal (SAR) is pegged to the US dollar . Hence, Saudi Arabia’s interest rate follows the US Federal Reserve interest rate and, therefore, monetary policy does not have much to do with economic growth, employment, and price stability.

The block contains definitional identities for all monetary aggregates, i.e., cash held outside banks, cash with demand deposits, a broad monetary aggregate, and a broad monetary aggregate with foreign currency deposits. It also has identities for interest rates, namely, the interest rate on lending, the effective interest rate on external debt, the interest rate on 10-year government bonds, and spread between the foreign and domestic interest rates. The block uses the definitional identity for the real exchange rate the real effective exchange rate. The latter one is determined by the ratio of domestic and foreign prices, together with the nominal effective SAR exchange rate against a basket of currencies of its main trading partners. This identity links the monetary block to the external and prices blocks. Real effective exchange rates are considered as a measurement of international price competitiveness in many theoretical and empirical studies. Therefore, we have econometrically modeled the impacts of the key theoretically predicted and Saudi-specific determinants, such as relative productivity in the oil and non-oil sectors, government consumption, net foreign assets on the real effective exchange rate, which allows us to calculate the equilibrium trajectory of the real effective exchange rate. (This equation is the slightly modified version of the equation estimated in Hasanov and Razek 2022) In this regard, the model can simulate how changes in domestic and foreign markets can shape Saudi Arabia’s international competitiveness. Such simulations might be important for policy making as SV2030 targets to increase Saudi Arabia’s position up to among the 15 most competitive economies over the world by 2030. Finally, the block contains another behavioral equation for broad money demand, M2 aggregate, which estimates it in real terms as a function of the interest rate differential between the domestic and foreign interest rates, GDP from the real sector, real oil price from the external sector, real effective exchange rate, and financial innovation proxied by time trend (This equation was borrowed from Hasanov et al. (2022) and slightly modified.) Section 7.3 provides details of the relationships of the monetary block.

6.4 External Block

The external sector in KGEMM is classified into exports, imports, and other balance of payment components.

Total exports are the sum of exports of goods and services. The exports of goods are broken down into oil exports and non-oil goods exports as stylized facts of the Saudi economy. The former is the sum of the exports of crude oil and oil refinery products. Crude oil exports are represented by an identity, which links it to domestic oil production minus domestic oil use, multiplied by the global price of Arabian light. It is one of the most important relationships in the block, as it links the external block to the real block, the energy block, and to the rest of the world. Thus, one can simulate the model to see how energy price reform could increase crude oil export revenues by lowering domestic oil use and freeing up more oil for export. Alternatively, one can look at the impact of international oil price dynamics and the impact of OPEC oil production agreements and exports on Saudi Arabia’s oil export revenues. The behavioral equation for the exports of refined oil products includes, alongside other variables, the world trade index for refined oil products. This index combines 45 individual countries’ demand for refined oil products. This enables the model to simulate how changes in the demand for refined oil products in a given country caused by renewable energy transitions or environmental policies influence Saudi Arabia’s refinery oil exports. The Kingdom’s oil exports feed into the government’s oil revenues in the fiscal block. The non-oil goods exports are estimated as being dependent on the real effective exchange rate, domestic production capacity, and demand from the rest of the world. Put differently, the equation brings together demand and supply-side factors and links the external sector to the real sector, and the monetary sector and the rest of the world. The total exports of services are the sum of the exports of services from the following nine activities: oil, investment income, other services, transportation, travel, communications, freight and insurance, financial services, and government. This detailed breakdown allows a modeler to simulate each service activity’s role in total exports and in overall economic performance.

Total imports are the sum of the imports of goods and services, broken down into the same categories as total exports. The imports of goods in turn broken into three categories, imports of capital goods, imports of intermediate goods, and imports of consumer goods. Each category is econometrically modeled as a function of domestic demand and the real effective exchange rate, which reflects the nominal effective exchange rate and differences in domestic and foreign prices. The imports of services are also estimated as a function of domestic demand and the real exchange rate. Such breakdown allows a modeler to examine the role of various imported goods and services in domestic demand which may provide insights about import substitutions and local content implications being important for the diversification of the Saudi economy. The block also contains overall and non-oil trade balance both determined as identities, that is, the difference between exports and imports as well as non-oil exports and imports, respectively.

As with other balance of payment components in the model, we have identities for foreign direct investments outflows and net foreign direct investment. Outflows are linked to the development of the Saudi economy and the net is the difference between the inflows and outflows. As mentioned above, the world trade index for refined oil products is represented by an identity combining 45 individual countries’ demand for refined oil products. KSA tourism demand indicator is also represented by an identity, which reflects the kingdom’s tourism demand for 10 countries, including developed and developing economies. Additionally, the block has an outflow of remittances econometrically estimated as a function of domestic economic activity measured by Saudi GDP, employed foreigners, living costs in Saudi Arabia measured by domestic price level, and expatriate levy. (The equation is borrowed from Javid and Hasanov 2022 and modified slightly.) This equation can provide insights about the main determinants of the outflow remittances, which is a leakage from the Kingdom and thereby diminishes the magnitude of its fiscal multipliers. Such insights are quite important given that fiscal policy-related economic development initiatives are dominant in the Saudi economy, and the Kingdom is one of the top countries globally in terms of outflow remittances. Lastly, the KGEMM team is examining the inflow of foreign direct investment using its theoretically predicted determinants such as productivity, macroeconomic stability, openness, and business costs, including the unit labor cost, infrastructure, and institutional development. This is the work under progress and not completed yet. Section 7.4 details the external sector relationships.

6.5 Domestic Prices Block

This block comprises three sub-blocks in a broader classification: consumer price indexes (CPIs), GDP deflators or producer price indexes, and aggregated energy prices. We discuss them briefly below.

The model has all 12 consumer basket prices as given by GaStat, estimated using behavioral equations. In forming the specifications for the CPIs we considered mainly supply-side factors using the markup and purchasing power parity approaches (e.g., see Brouwer and Ericsson 1998; Juselius 1992). This is because we assume that Saudi inflation is primarily cost-push inflation. The markup approach factors we considered were unit labor cost or just wages, domestic energy prices, producer prices, and VAT rate. In this framework, a modeler to simulate the effects of policy interested variables (such as  VAT rate, domestic energy prices) and foreign prices on domestic consumer prices. As the Saudi economy undergoes a transformation process in line with SV2030, future versions of the KGEMM could also incorporate the money market and output gap approaches in inflation modeling. In addition to the 12 estimated equations, the sub-block contains an identity for overall CPI, which is a weighted average of the 12 CPI components. The weight variable of each of 12 components contains two values, old weights from 1970 to 2012 and new weights for 2013–2019 as documented in Appendix E.

GDP deflators, considered as the producer prices, sub-block, estimates GDP deflators for economic activity sectors, considering the respective overall price (i.e., deflators for non-oil GDP, oil GDP, and GDP) as a catch-up factor and domestic energy prices, foreign/import prices among other control variables. This setup allows a modeler to simulate the model to investigate the impact of the ongoing energy price reforms on the production costs, and thus the competitiveness of the economic activities.

The aggregated energy prices sub-block mainly contains weighted average domestic prices of energy products by customer type for several economic activity sectors, if a given sector uses more than one energy product. For example, the aggregate energy price for the utility sector is the weighted average price of natural gas, crude oil, diesel, and heavy fuel oil. The weighted average domestic energy price was calculated for the following economic activity sector as well: distribution, agriculture, financial and banking services, other services, transportation and communication, construction, and government services. Such a weighted average price allows us to simulate the model for both the price and demand effects of different energy products on a given economic activity sector under consideration. It should be noted, however, that with the exception of the utility sector (and electricity consumption to some extent), we do not have data on the consumption of energy products in the economic activity sectors. Therefore, the calculated weighted average prices for the economic activity sectors other than the utility sector are only rough approximations. A detailed description of the domestic prices block is documented in Sect. 7.5. 

6.6 Labor and Wages Block

This block contains relationships for employment, wages, and unit labor cost for economic activity sectors. The block also comprises definitional identities for labor force, unemployment, and its rate.

Total employment is broken into oil sector and non-oil sector employment. Employment in the oil sector is the sum of employment of oil mining and oil refinery. Employment in the non-oil sector is the sum of employment in 11 non-oil economic activities. Employment of each non-oil economic activity is econometrically estimated as a function of wage and output in a given activity sector (see Hasanov et al. 2021). We also have employment in the private sector and it is the sum of nationals and foreigners.

The wage equations for some economic activity sectors are econometrically estimated using output price and labor productivity as main determinants. However, they have not yet been completed for all sectors at the time of writing and, hence, cover seven economic activity sectors. Employment and wage equations establish links with the real block and domestic prices block.

The block contains identities that represent the unit labor costs for 12 economic activity sectors (and three aggregated sectors, that is, service, oil and non-oil). The identities use the conventional definition for unit labor cost, that is, each sector’s unit labor cost is the sector-specific average wage rate multiplied by the sector-specific employment divided by the sector-specific value added. Also, the block has economy-wide labor compensation. This is the sum of labor compensation (wage rate times employment) in 11 non-oil activities and two oil activities. This variable links this block to the real block of the model as the variable feeds into the identity for disposable income. In addition, it provides the capability to simulate the model for the impact of different wage rates and employment in various sectors on households’ disposable income.

Lastly, considering the stylized facts of the Saudi economy, the labor force is linked to working age population groups of Saudis and non-Saudis, both broken into the sum of males and females coming from the population and age cohorts block. This allows one to differentiate the role of each group and their male and female components in the labor force and, thus, also in the unemployment. A detailed description of the labor market and wages block is documented in Sect. 7.6.

6.7 Energy Block

The energy block differentiates KGEMM from conventional (semi-)structural macroeconomic models. The block comprises demand for different energy products in volume and value (monetary) measures as well as the supply of electricity.

The block has econometrically estimated 15 behavioral equations for energy demand in volume terms. As a volume term, we use million tons of oil equivalent (MTOE) for all the energy products to make them comparable with each other. There are nine energy products (crude oil, diesel, heavy fuel oil [HFO], natural gas, electricity, liquefied petroleum gas [LPG], kerosene, gasoline, and other oil products) and six customer types (residential, industry, commercial, government, transport, agriculture and forestry). In addition to this, the first four energy products are used in the utility sector, but they are not econometrically estimated as the sector is mostly government-owned, that is, exogenous. However, not all six customers consume all 15 energy products (e.g., electricity is consumed by all customers except transport, whereas only industry and utility consume HFO and crude oil). The equations have been estimated using the conventional energy demand framework, where demand for a given energy product is a function of its own price and customer-specific income, both in real terms. For some energy products, we extended the conventional framework with other explanatory variables as deemed reasonable. For instance, we had real prices of substitutable energy products, where they became theoretically interpretable and statistically significant alongside own price and income in the estimations (see estimated equations for the industrial demand for crude oil, HFO, natural gas; for the transport demand for diesel; for the residential demand for kerosene and LPG). Also, we included cooling degree days in the residential electricity demand equation and accounted for the population effect. The latter effect was also accounted for in the gasoline demand estimation. Moreover, we had a working age group population in the industrial electricity demand estimations following the theoretical framework developed in Hasanov et al. (2021). Furthermore, we found that the investments are statistically significant and have a negative sign, indicating efficiency gains in the electricity demand equations for commercial and agricultural sectors. The customer-specific price deflators have been used to calculate energy prices in real terms, as suggested by the energy demand literature. The estimated equations mainly link this block to the domestic prices block and real block. Recall that energy consumption also feeds into the demand-side estimation of the economic activities in the real block.

The block also represents the above-mentioned 15 energy products consumption in value terms, that is, their volumes in MTOE are multiplied by their prices in Saudi riyals for per TOE. This monetary representation is for the purpose of calculating government energy sale revenues by product, customer, and by total economy. Recall that total energy consumption in monetary terms feeds into the government’s non-oil revenues in the fiscal block.

The supply of electricity is broken into two generation sources: fossil fuels and renewable. Fossil fuels-based generation accounts for four main energy products used in the electricity generation in Saudi Arabia, namely, crude oil, diesel, HFO, and natural gas. The sum of these fossil fuels is multiplied by the efficiency ratio to calculate the amount of electricity generated. The second part of the electricity generation is coming from solar energy sources. We did not consider renewable energy sources other than solar, as they are very negligible historically. Even solar power generation has an average share of 0.1% of total electricity generation over the 2010–2021 period. The good news is that this share has increased significantly in recent years. Saudi Arabia plans to increase the share of renewables in total electricity generation demanded to 50% by 2030, with the other 50% coming from natural gas. This implies excluding the other fossil fuels from the power generation. Having total electricity generation coming from fossil fuels and solar sources makes KGEMM unique in simulating the economic, energy, and environmental effects of different scenarios for displacing fuels with renewables as it was done in Elshurafa et al. (2022) and Hasanov et al. (2022a). The above discussed behavioral equations and identities in addition to others are represented in Sect. 7.7. 

6.8 CO2 Emissions Block

This is another block that makes KGEMM different from conventional (semi-) structural macroeconometric models. The literature on environmental pollution shows that usually, energy-related CO2 emissions are around 90% of the total CO2 emissions. Therefore, this block was constructed using the energy block above. It is calculated as volume of a given energy product consumed (e.g., crude oil) multiplied by the product-specific conversion factor to reach up to the amount of CO2 emissions. We used each product-specific emission conversion factor in the calculation. The conversion factors are retrieved from various reputed sources such as the International Energy Agency, US Energy Information Administration, US Environmental Protection Agency. Appendix E documents conversion factors and their sources while Sect. 7.8 reports calculations, i.e., constructed identities. We grouped product-based CO2 emissions by customer type—industry, transport, residential, commercial, government, agriculture, but other classifications can be considered as well. As mentioned at the beginning of this section, CO2 emissions are one of the new developments in the fifth version of KGEMM meaning that it has room for extension. We plan to expand the block by incorporating other emission indicators into it. For example, a carbon price or carbon tax is one of the considerations for future work. The block allows a modeler to assess the impacts of different CO2 emissions reduction options on energy consumption and economic indicators.

6.9 Population and Age Cohorts Block

The total population in the block is the sum of the 12 age groups: 0–14, 15–19, 20–24, 25–29, 30–34, 35–39, 40–44, 45–49, 50–54, 55–59, 60–64, 65 and above. Each age cohort is determined as the sum of males and females. The block represents the total Saudi population as the sum of Saudi males and females. The same formula is applied to the non-Saudi population. The working age population is the sum of males and females. Both male and female working age population groups are the sum of age cohorts 15–19 through 60–64. The population of Saudis and non-Saudis feeds into the labor force in the wages and labor block as mentioned above. One can simulate the model to assess the impacts of changes in each of the 12 age cohorts broken into females and males on the on various relationships and the total economy, as well as the Saudi and non-Saudi impacts on the labor force and unemployment. Details of the block can be found in Sect. 7.9