Climate Change And Economic Growth: An Intertemporal .

6m ago
274.44 KB
22 Pages

Climate change and economic growth: An intertemporal generalequilibrium analysis for EgyptAbeer Elshennawya, Sherman Robinsonb, Dirk Willenbockelc,*aAmerican University in Cairo, New Cairo 11835, P.O. Box 74, [email protected] Food Policy Research Institute, 2033 K St NW, Washington, DC20006-1002, USA. [email protected] of Development Studies at the University of Sussex, Brighton BN1 9RE, [email protected]*Corresponding author: Institute of Development Studies at the University of Sussex, Library Road,Brighton BN1 9RE, UK. Tel: 44 1273 915700.Abstract:This study develops a multisectoral intertemporal general equilibrium model with forward-lookingagents, population growth and technical progress to analyse the long-run growth prospects for Egypt ina changing climate. Based on a review of existing estimates of climate change impacts on agriculturalproductivity, labor productivity and the potential losses due to sea-level rise for the country, the modelis used to simulate the effects of climate change on aggregate consumption, investment and welfare upto 2050. Available cost estimates for adaptation investments are employed to explore adaptationstrategies.The simulation analysis suggests that in the absence of policy-led adaptation investments, real GDPtowards the middle of the century will be nearly 10 percent lower than in a hypothetical baselinewithout climate change. A combination of adaptation measures, that include coastal protectioninvestments for vulnerable sections along the low-lying Nile delta, support for changes in cropmanagement practices and investments to raise irrigation efficiency, could reduce the GDP loss in 2050to around 4 percent.JEL Codes: C68, D58, D90, E17, O44, Q54Keywords: Climate change adaptation; Computable general equilibrium analysis; Scenario analysis;Dynamic CGEPaper for 17th Annual Conference on Global Economic Analysis, Dakar (Senegal),June 20141

1. IntroductionDue to the high concentration of economic activity along the low-lying coastal zoneof the Nile delta and its dependence on Nile river streamflow, Egypt's economy ishighly exposed to adverse climate change. Adaptation planning requires a forwardlooking assessment of climate change impacts on economic performance at economywide and sectoral level and a cost-benefit assessment of conceivable adaptationinvestments.This study develops a multisectoral intertemporal general equilibrium model withforward-looking agents, population growth and technical progress to analyse the longrun growth prospects of Egypt in a changing climate. Based on a review of existingestimates of climate change impacts on agricultural productivity, labor productivityand the potential losses due to sea-level rise for the country, the model is used tosimulate the effects of climate change on aggregate consumption, investment andwelfare up to 2050. Available cost estimates for adaptation investments are employedto explore adaptation strategies.On the methodological side, the present study overcomes a basic limitation of existingcountry-level recursive-dynamic computable general equilibrium models 1 for climatechange impact analysis by incorporating forward-looking expectations. In contrast tothe standard recursive-dynamic approach, in which climate shocks hit agents in themodel by surprise, the intertemporal approach pursued here takes account ofendogenous anticipative adaptation responses to expected future climate changeimpacts. Moreover, it extends the existing family of discrete-time intertemporalcomputable general equilibrium models to which our model belongs by incorporatingpopulation growth and technical progress. On the empirical side, the model iscalibrated to a social accounting matrix that reflects the observed current structure ofthe Egyptian economy, and the climate change impact and adaptation scenarios areinformed by a close review of existing quantitative estimates for the size order ofimpacts and the costs of adaptation measures.1Examples for recent country-level studies using recursive-dynamic CGE models include Arndt et al(2011,2012), Robinson et al (2012) and Thurlow et al (2012). For an early study of this type for Egyptsee Strzepek and Yates (2000). Fankhauser and Tol (2005) and Lecocq and Shalizi (2007) providesystematic conceptual discussions of the channels through which climate change potentially affectsaggregate economic growth in Solow-type growth models, Cass-Koopmans-type optimal growthmodels and endogeneous growth models. Babiker et al (2009) compare recursive-dynamic andintertemporal specifications in global climate change mitigation modeling.2

The following section outlines the model and its numerical calibration. Section 3specifies and motivates the climate change impact simulation scenarios. Section 4presents simulation results in the absence of policy-led adaptation investments.Section 5 considers adaptation scenarios, section 6 reflects briefly on sensitivity andlimitations of the analysis, and section 7 concludes.2. The modelThe determination of intertemporal saving and investment decisions in the model isessentially a multi-sector open-economy extension of neoclassical optimal growththeory in the Ramsey-Cass-Koopmans tradition, while intratemporal allocationdecisions across sectors are determined by a standard static small open economy CGEmodel as described in full technical detail in Robinson et al (1999). The operationalmodel design draws upon the contributions to intertemporal CGE analysis and itsapplications by Go (1994), Mercenier and Sampaio de Souza (1994), Diao andSomwaru (1997), Elshennawy (2011) and Roe et al. (2010), but extends this class ofapplied models by incorporating population growth and technical progress.In line with its theoretical pedigree, the long-run steady-state growth rate of the modelis governed by labor force growth and the rate of technical progress, while climateimpacts that affect savings and investment entail level shifts in the time paths of GDP,consumption and other macroeconomic aggregates without affecting the long-runtrend growth rate.For purposes of the present study, the model distinguishes six sectors of economicactivity: agriculture, oil, industry, construction, electricity and services. Output isproduced using intermediate inputs and primary factors of production which includelabor and capital. To capture the impact of different policy scenarios on the labormarket, two skill categories of labor are distinguished, production and nonproductionlabor. For simplicity, the role of government is confined to tax collection. Taxrevenue is redistributed to the household sector and government expenditure is treatedas part of household consumption. The agents in the model are a representativehousehold with infinite planning horizon, a representative firm in each of theproduction sectors, and the rest of the world, which is linked to the domestic economyvia trade, transfer and capital flows. Markets are perfectly competitive. What followsis a description of the dynamic components of the model.3

2.1. Consumption behaviorThe representative household receives labor and dividend income from firms as wellas net transfer income from the rest of the world and the re-transfer of tax revenue.The household chooses the path of consumption that maximizes the intertemporalutility function(1)𝐢1𝑑subject to the intertemporal budget constraint(2)𝐢1 𝑛 𝑑𝑑𝑑 π‘ˆπ‘œ 𝑑 0 𝑁𝑑 ln 𝑁 (1 𝜌)𝑑 𝑁0 𝑑 0 𝑙𝑛 𝑁 1 𝜌 𝑑 𝑑 0 𝑅𝑑 𝑃𝑑 𝐢𝑑 𝑑 0 𝑅𝑑 [𝑀𝑝𝑑 𝐿𝑃𝑑 𝑀𝑛𝑑 𝐿𝑁𝑑 𝑇𝑅𝑑 𝑇𝑋𝑑 ] π‘Š0and a no-Ponzi-game transversality condition, where C is an index of aggregate realconsumption, N LP NP is household size with LP and NP denoting productionand non-production labor respectively, n is the rate of population and labor forcegrowth, ρ is the pure rate of time preference, P is the implicit consumer price indexdual to C, wp and wn are the wage rates for production and non-production labor, TRdenotes net transfer income from the rest of the world, TX is tax revenue, W0 is initialfinancial net wealth of the household sector, which is equal to the total market valueof the firms owned by the representative household minus the initial external debtowed to the rest of the world, and(3)𝑅𝑑 𝑑𝑠 0 1/(1 π‘Ÿπ‘  )is the discount factor where r denotes the world interest rate.The first-order conditions for the maximization of (1) subject to (2) and thetransversality condition, which ensures that the given initial debt does not exceed thepresent value of future current account surpluses, take the form(4)𝑃𝑑 1 𝐢𝑑 1 1 πœŒπ‘ƒπ‘‘ 𝐢𝑑1 𝑛 1 π‘Ÿπ‘‘ .2.2. Investment behaviorIn each model sector s, firms are aggregated into one representative firm whichfinances all of its investment through retained earnings and thus the number of equitiyshares issued remains constant. Managers seek to maximize the value of the firm.Assuming perfect capital markets, asset market equilibrium requires equal rates ofreturns (adjusted for risk) on all assets. This implies that firm’s equity must earn an4

expected rate of return equal to that of a safe asset like foreign bonds as reflected inthe condition(5)r DIVs Vs VsVswhere DIV is dividends, V is the value of the firm, Vs Vs,t - Vs,t-1 is the expectedannual capital gain on firm equity and r is the interest rate on foreign bonds.Solving the above difference equation (5) forward yields(6)𝑉𝑑 𝑣 𝑑 𝑅𝑑 𝐷𝐼𝑉𝑑 .The market value of the firm equals the discounted stream of future dividends.Dividends distributed to the household sector equal operating surplus minusinvestment expenditure:(7) 𝐷𝐼𝑉𝑆,𝑑 𝑃𝑉𝐴𝑆,𝑑 𝑓 𝑏𝐿𝑃𝑆,𝑑 , 𝑏𝐿𝑁𝑆,𝑑, 𝐾𝑆,𝑑 𝑀𝑝𝑑 𝐿𝑃𝑆,𝑑 𝑀𝑛𝑑 𝐿𝑁𝑆,𝑑 𝑃𝐼𝑆,𝑑 𝐼𝑑 𝐴𝐷𝐢𝑆,𝑑 ,where, f (.) is the production function, K is capital, PI is the price per unit ofinvestment I, PVA is the value added price (output price net of indirect productiontaxes and intermediate input unit costs) and ADC represents adjustment costsassociated with the installation of new capital:(8)𝐼2𝐴𝐷𝐢𝑆,𝑑 𝑃𝐼𝐴𝑆,𝑑 πœ‘ 𝐾𝑆,𝑑𝑆,𝑑Due to the presence of these adjustment costs, the capital stock does not adjustinstantaneously to its new optimal long-run level following exogenous shocks thataffect the return to capital. Adjustment costs to investment are assumed to be internalto the firm. For any given level of the capital stock these costs are strictly increasingin investment and decreasing in the capital stock for any given level of investment.As a result, firms will find it optimal to increase the capital stock gradually over timein order to reach the optimal long run capital intensity. The adjustment cost functionis assumed to be linear-homogeneous in investment and capital. Along with theassumption of constant returns to scale in production, the linear homogeneity of theadjustment cost function entails that Tobin’s marginal q equals Tobin’s average q(Hayashi, 1982). In the general equilibrium model, the real adjustment costs take theform of purchases of installation services, which are a Leontief composite of theconstruction and industry commodities, and PIA is the unit price of this composite.The model incorporates labor-augmenting technical progress. The labor efficiencyparameter b in (7) grows at the uniform exogenous rate g.5

In each sector producers maximize the value of the firm subject to the capitalaccumulation constraint(9)𝐾𝑆,𝑑 1 (1 𝛿𝑆 )𝐾𝑆,𝑑 𝐼𝑆,𝑑 ,where Ξ΄ is the rate of depreciation. Differentiating the Lagrangean for thisoptimization with respect to the control variable I yields(10)πΌπ‘žπ‘†,𝑑 𝑃𝐼𝑆,𝑑 2𝑃𝐼𝐴𝑆,𝑑 πœ‘ 𝐾𝑆,𝑑 ,𝑆,𝑑which determines the shadow price of capital. Condition (10) states that the firminvests until the cost of acquiring capital – which is equal to the price of a unit ofinvestment plus marginal adjustment costs– is equal to the value of capital.Differentiating with respect to the state variable K yields the no arbitrage condition(11)𝐼2𝑃𝑉𝐴𝑆,𝑑 𝑓𝐾 𝑃𝐼𝐴𝑆,𝑑 πœ‘ 𝐾𝑆,𝑑 (1 𝛿)π‘žπ‘†,𝑑 (1 π‘Ÿ)π‘žπ‘†,𝑑 1 0 .𝑆,𝑑According to Equation (11), the value of the marginal product of capital PVA fK plusthe marginal reduction in adjustment costs brought by the increase in capital plus thecapital gains qt - qt-1minus depreciation Ξ΄q must equal the amount foregone rq bychoosing to accumulatethis extra unit of capital. For simplicity, there is nodifferentiation between government and private investment in the model. IS,t is aCobb-Douglas composite good over commodity groups demanded for investmentpurposes,(12)ΞΈI S ,t AK S S / INVDS S/ ,/S,S,where INVDS’,S is investment demand by sector S for goods of type S’ and AKS is aconstant parameter. PIS,t is the investment price index dual to IS,t .2.3. Current account dynamicsThe current account dynamics associated with the optimal consumption andinvestment path is described by(13)𝐷𝑑 1 𝐷𝑑 π‘Ÿπ‘‘ 𝐷𝑑 𝑇𝐡𝑑 π‘‡π‘…π‘‚π‘Šπ‘‘ ,where TBt is the trade balance surplus in t and TROW denotes exogenous nettransfers from abroad. Letting Y denote aggregate GDP, TBt Yt - PtCt - S PIS,tIS,t.The no-Ponzi-game condition invoked in the derivation of the optimal consumptionpath described by (4) entails that the initial debt inherited from the path constrains thefuture path of domestic absorption, so that D0 PV(Yt TROWt) – PV(PtCt) – PV( SPIS,tIS,t), where PV(x) denotes the present value of a stream xt discounted at rate r. In6

other words, the initial debt must be matched by a corresponding positive presentvalue of future primary account surpluses.2.4. Intratemporal general equilibriumEmbedded in this dynamic structure is a standard within-period general equilibriummodel that determines intratemporal relative prices, the sectoral allocation of laborand the commodity composition of consumption, imports and exports.Producers in the model are price takers in output and input markets and use constantreturns to scale technologies described by constant elasticity of substitution (CES)value added functions and a Leontief fixed-coefficient technology for intermediateinput requirements by commodity gro