Climate Change and Economic Growth: An Integrated Approach to Production, Energy, Emissions, Distributions and Unemployment

Transcription

1 Climate Change and Economic Growth: An Integrated Approach to Production, Energy, Emissions, Distributions and Unemployment Asjad Naqvi September 3, 2014 PRELIMINARY DRAFT Abstract This paper presents the ECOGRO model that combines production, consumption, and emissions in a stock-ow consistent (SFC) macro framework. The model introduces four innovations. First, it vertically integrates the energy sector with the rm sector to allow for feedback mechanisms within the supply chain in a demand driven framework. Second, energy is supplied by two energy producing agents; a non-renewable input dependent, high emissions energy producer and a capital intensive, renewable energy producer. Third, capital and energy eciency parameters are introduced where dierent rates of innovation give insights on sustainable growth paths with zero or decreasing emissions. Fourth, the environment is introduced as a sector that accumulates Greenhouse Gasses (GHGs) which endogenously feed back on the economy through taxes pegged to the level of emissions. Keywords: stock-ow consistent, production, energy, emissions, distributions, employment JEL: E12, E17, E23, E24, Q52, Q56 This project is funded under the European Commissions 7th Framework grant titled Wealth, Welfare, Work for Europe. The paper greatly beneted from preliminary conversations with Antoine Godin, Stephen Kinsella, Kurt Kratena, Armon Rezai, Miriam Rehm, Sigrid Stagl, the participants at the WWW4Europe Modeling Workshop and the University of Limerick Workshop on Evironmental Modeling. All errors are my own. Post-Doc, Institute for Ecological Economics, Department of Socioeconomics, Vienna University of Economics and Business (WU), Austria. snaqvi@wu.ac.at. 1

2 1 Introduction The paper presents a demand-driven multi-sector macro model that inter-links production, energy, and the environment in a stock-ow consistent (SFC) framework (Taylor (2004); Godley and Lavoie (2007); Santos and Zezza (2008); Caverzasi and Godin (2013)). The main motivation behind this paper is to address the concerns raised in recent literature that highlight the need to recognize the interdependence of economic activity and the environment (Jackson (2009); Spash (2012); Rezai, Taylor, and Mechler (2013); Victor and Jackson (2013); Fontana and Sawyer (2013)). The reason for this is two-fold. First, economy activity is complex where institutional sectors of the economy (households, rms, nancial sector, government) update their decision making process based on the actions of others.. Thus any climate related policy cannot be studied in a vacuum because its impact on all social actors of an economic system have to be discussed especially if feedback mechanisms result in unintended consequences towards policy outcomes. This is especially crucial for analysis in the current economic state when economies have stagnated and growth is extremely low (European Commission (2014)). This coupled with concerns over income and wealth distributions (Piketty (2014)) and unemployment (Stockhammer (2004)) raises huge implications with policy suggestions that include raising taxes or pursuing persistent low growth regimes (Jackson (2009); Victor (2012)). The modeling framework used in this paper is based on the stock-ow consistent (SFC) accounting framework developed in Godley and Lavoie (2007). The SFC framework represents a closed monetary economy where dierent institutional sectors - households, production sector, nancial institutes, and the government - interact endogenously to generate economic outputs (Taylor (2004); Godley and Lavoie (2007); Santos and Zezza (2008)). The framework follows the national accounts method of using a balance sheets to represent stocks or net worth of the economy and a transition matrix to represent inter-institutional ows. A key advantage of this framework is that all interactions and transactions are tracked across all sectors making the model fully tractable. In order to avoid the complexity of representing a fully functioning economy the model only develops the production sector in detail which integrates rms vertically with the energy sector, a key input in the production process (Kronenberg (2010); Fontana and Sawyer (2013)). The monetary economy is integrated with the environment through two dierent channels (Jackson (2009); Fontana and Sawyer (2013); Taylor and Foley (2014)). First is the cost of extraction of a non-renewable resource that is required to produce energy. As the stock of the non-renewable resource depletes, input prices go up directly aecting the cost of the output produced (Victor (2008)). Second, the production process results in emissions that can have an adverse eect on the economy through a damage function. While this damage function has been studied in detail mostly from a quantiable angle in term of output lost (Nordhaus (1992); Stern (2007); Foley, Taylor, and Rezai (2013); Pindyck (2013); Taylor and Foley (2014)), in this paper emissions are tackled from a dierent angle. Scientists unanimously agree on the negative extesrnalities of emissions (IPCC (2007, 2012)) but free market options, for example emissions trading regimes in the EU (Laing, Sato, Grubb, and Comberti (2013)) have mostly been unable to curb rising emissions from human activity. In order to address this concern, active regulatory policies to mitigate emissions are discussed. While all policies discussed in the paper are aimed at reducing emissions in the most eective manner, the paper also focuses on discussing the impact of 2

3 policies on pricing mechanisms, income distributions, and unemployment levels. The paper is organized as follows. Section 2 sets up the framework for the ECOGRO model and Section 3 explains the relationship in detail. Section 4 describes the simulation results for four dierent policy experiments. Section 5 concludes. 2 Framework Figure 2.1 summarizes the relationships between the key sectors in the model. The production sector at the core of the model comprises rms and the energy producers. Households interact with the production sector through capitalists who own the production sector and workers that are employed in the rms. Households and the production sector interact with the nancial sector through deposits and loans. The government earns income through taxes from the production, household, and nancial sector which it uses to nance its consumption and support the unemployed. The environment both provides the non-renewable input for producing energy and accumulated Greenhouse Gasses (GHGs). Figure 2.1: Model layout The model makes several key assumptions about the economy. It is assumed to be demand- 3

4 driven where a rise in real income levels makes the economy grow and reduces the unemployment levels. The model show the discusses distribution between two broad types of households; capital owners and workers. Capital owners are assumed to hold productive capital which is operated by workers in exchange for wages. These are determined exogenously in a bargaining process between the two household classes. Capitalists earn dividends and prots from the rms and the nancial sector. The production function takes a Leontief form with three inputs; capital, labor and energy as perfect compliments. Energy is assumed to be only consumed by rms. Firms set prices as markup over costs that then determine the real variables in the system, including the household demand and the unemployment level. The burden of any rm level tax is shifted to consumers through prices. Agents hold stocks, and economic activity results in ows between agents. Following the system developed in Godley and Lavoie (2007), Table 2.1 shows the balance sheet of the model and Table 2.2 shows the TFM. Table 2.1: Balance Sheet Households Production Financial Govt. Unemp. Workers Capitalists Firms Energy - X Energy - R Banks Central Bank Capital stock +K +K X +K R +K Inventories +IN +IN X +INV Cash +M h +M k +M 0 Bank Deposits +D h +D k D b 0 Advances A b A 0 Bills +B b +B CB B 0 Loans L f L X L R +L 0 0 +V h +V k +V f +V X +V R 0 0 V G +NV The columns in the balance sheet gives the net worth of agents while the rows represent the net-worth of assets. The asset prole of the household sector is represented by cash and deposits while the production sector is valued in terms of capital stock, inventories and loans. Commercial banks hold deposits against which loans are given out while also investing in treasury bills issued by the government to hedge against changes in deposits. If needed, this investment is made by taking advances from the central bank which acts as a lender of last resort. The central bank issues cash and hold any treasury bills not held by the private sector. The government only has liabilities which are the treasury bills that it issues to nance any decit. At the aggregate level, the households and the production sector has a positive network while the government has a negative net worth. Flows resulting from the interaction between agents generate the changes in balance sheets between time periods. These ows are summarized in the TFM given in Table 2.2. Accounting restrictions imply that that all rows and columns must add up to zero. Columns represent the sources and uses of funds for each agent category. For example, the workers column shows wages and interest earnings on deposits as inows while taxes and consumption are outows. Any savings after expenditure are allocated to either cash or bank deposits which are also reected in the balance sheet. Rows represent the ows for each line item. For example, the rst row shows that total goods purchased for consumption by households and the government should match the total sales of the rms. 4

5 Both the balance sheet and the TFM show the values in nominal terms. Section 3 describes these interactions in detail. 5

6 Unemp. Workers Capitalists Table 2.2: Transition Flow Matrix Firms Energy - Dirty Energy - Clean Commercial Banks Central Bank Current Capital Current Capital Current Capital Current Capital Current Capital Consumption C u C h C k +S G 0 Energy EB +E X +E R 0 Investment + I I + I X I X + I R I R 0 Inventories + IN IN + IN X IN X 0 Wages +W B W B 0 Unemp. Benets +UB UB 0 Bank prots +Π b Π b 0 Firm prots +Π f Π f 0 Energy prots +Π E Π X Π R 0 CB prots Π CB +Π CB 0 Taxes T h T k T f T X T R +T 0 i Advances raat 1 +raat 1 0 i Deposits +rdd h t 1 +rdd k t 1 rddt 1 0 i Bills +rbb b t 1 +rbb CB t 1 rbbt 1 0 i Loans rll f t 1 rll X t 1 rll R t 1 +rllt 1 0 Advances + A A 0 Cash C h C k + C 0 Deposits D h D k + D 0 Bills + B b + B CB B 0 Loans + L f + L X + L R L Govt. 6

7 3 Model This section describes the behavioral equations of agents dened in two categories. Behavioral equations, which are derived from literature and dene the decision making process of agents for example how consumption is dened or how investment decisions are made. Identities that must hold based on accounting principles. For example, savings have to equal income less expenditures. Behavioral decisions are made in real terms, while identities are calculated in nominal values as residuals of the decision making process. In order to make a distinction between the two, capital letters are used to represent nominal (current) value dened in monetary terms while lowercase letters represent real values or stocks. Nominal variables are converted into real variables using the endogenously determined price level p t. For dierent agent categories, variables are super-scripted using h for workers, k for capitalists, u for unemployed, f for rms, X for the non-renewable energy sector, R for renewable energy sector, b for commercial banks, CB for the central bank, and g for government. Time is denoted with a subscript t and exogenous parameters are written using Greek symbols. represents a rst order dierence such that x t = x t x t 1 The subsections below describe the key equations; while the full model is summarized in Appendix??. 3.1 Firms The total output in the economy y t equals total sales s t plus changes in inventories in t (3.1). y t = s t + in e t in t 1 (3.1) s t = c k t + c h t + c u t + Γ (3.2) Assuming agents have full information about current demand, total sales (3.2) equal total household consumption plus an exogenously dened government consumption level Γ. N f t = y t ξ Y N (3.3) W B t = N f t ω (3.4) The total number of workers hired by the rm N f t equals total output produced each time period divided by the exogenously dened labor productivity per unit of output ξ Y N. The wage bill is estimated as total workers hired times the exogenous wage rate ω. E f t = y t ξ Y E (3.5) EB t = E f t.p E t (3.6) 7

8 Firms require energy to produce the output y t where the energy demand in units E f t equals total output divided by the exogenous output to energy productivity ratio ξ Y E (3.5). The total energy bill (3.6) equals energy demand times the endogenously determined energy price p E t. UC t = W B t + EB t y t (3.7) p t = UC t (1 + θ t )(1 + τ F ) (3.8) 3.7 gives the rms unit cost which is derived as a simple average of total costs over total output. 3.8 highlights the rms pricing decision. Prices are determined by an endogenous markup θ t and the tax rate τ F which rms shift onto consumers through the prices. An endogenous markup is introduced here to allow for smoother adjustment of prices based on prot and dividend targeting. While the long-run equilibria are qualitatively the same with a xed and an endogenous markup, the latter allows for cyclicality in the adjustment process. The endogenous markup θ t is calculated as: θ t = Where Π T is the target prot formally derived as: Π T s t 1 UC t 1 (3.9) Π T t = D t + W B t 1 + EB t 1 + r l L f t 1 (3.10) (in T t in t 1 ).UC t 1 s t 1.UC t 1 that ensures the target level of dividends D t : D t = ι D Π t 1 (3.11) Here, ι D is the exogenous ratio of past prots Π t 1 that capitalists wish to earn as income. Firms' realized prots thus equal Π f t = S t (1 τ F ) + IN t W B t EB t (r l L f t 1 ) (3.12) or revenues which in this case is nominal sales S t = s t p t less taxes, production costs and interest payments. These prots are fully redistributed to capitalists. Firms make two investment decisions every time period that determine its demand for loans; a decision to invest in inventories and a decision to invest in capital stock. 8

9 in e t = in t 1 + γ(in T t in t 1 ) (3.13) in T t = σs t 1 (3.14) in t = y t s t (3.15) IN t = in t.uc t (3.16) Inventories are held by rms to hedge against unexpected changes in demand. The target level of inventories in T is determined through an exogenous inventories-to-sales target σ (3.14). The desired increase in inventories every time period in e t (3.13) equals a fraction γ of the gap between target inventories in T t and past inventories in t 1. The realized change in inventories in t at the end of each time period equals output less sales (3.15) gives the nominal value of inventories determined at unit cost UC t. In a full information model in e t in t will always hold. The rms' decision to invest in capital stock is determined by a desired level of capacity utilization ratio ν. Firms, as part of their liquidity preference strategy, wish to hold a certain proportion of their capital stock slack in order to adjust to future changes in demand. Capital stock growth (3.17) is determined as a fraction β of the gap between current capacity utilization rate u t and target capacity utilization ν. k t = k t 1 + β(u t ν)k t 1 (3.17) Current capacity utilization (3.18) is described as the current output divided by maximum potential output ȳ t. u t = y t ȳ t (3.18) ȳ t = ξ Y K k t 1 (3.19) ȳ t is determined by the exogenous capital to output parameter ξ Y K. An increase in the value of ξ Y K implies that a lower amount of capital is required to produce the same level of output y. Investment in capital is derived as i t = Max[β(u t ν) + δ, 0]k t 1 (3.20) I t = i t.p t (3.21) The value of the investment i t (3.20) is a non-negative function which equals target increase in capital stock plus depreciation δ. In other words, negative investment is not permitted capitalists will at least invest to balance. However, if the demand and subsequently production falls, then 9

10 capitalists will let their capital stock depreciate such that the value of the capital stock goes down give the nominal value investment respectively. Given that the rms are assumed to be fully leveraged, the total value of loans requested by rms from the banks equals the desired level of change in the nominal value of capital stock and inventories: L f t = I t + IN t (3.22) 3.2 Energy Sector The energy sector consists of two energy producing agents that provide a uniform energy input to rms; a high emissions energy producer that requires a non-renewable input X, and a capital intensive, low emissions energy producer that requires a renewable resource R. The energy sector follows the same production and investment decision as rms with three key dierences described below. First, the quantity of energy produced by each The total demand for energy is derived in 3.5 as a function of total output and the energy to output ratio ξ Y E. This demand is met by the two energy producing agents based on an exogenously dened share φ X of non-renewable energy and (1 φ X ) of renewable energy. The high emissions producer requires a non-renewable input X to produce energy. Assuming energy cannot be stored, rms hold inventories of non-renewable input X to smooth out future changes in demand. The total sales of X required to meet allocated energy demand are given as: s X t = φ XE f t ξ XE (3.23) The numerator gives the share of current energy demand that needs to be met by the highemissions producer and the numerator gives the energy to input X production ratio ξ XE. A higher value of energy productivity ξ XE thus implies lower requirement of input X. Following the same logic as for rms, total output of X generated by producer X is given as: y X t = s X t + (in XT t in X t 1) (3.24) or the total sales plus changes in inventories of X. The high emissions energy producer faces only the extraction cost of X which is derived as: XC t = κ X XIF t yt X (3.25) XIF t = 1 1 (yt X / X t ) (3.26) 10

11 where κ X is the unit cost of extraction and XIF t is the extraction ination factor. XIF t gives the rate at which extraction costs increase as the current stock of the non-renewable input X t is depleted over time. If a very small fraction of total output is extracted up to a time period t, such that the ratio in the denominator is close to 0, then XIF t will be close to 1 and will have a negligible impact on costs. The unit cost (3.27) is simply calculated as the total extraction costs over the total output produced. UC X t = XC t y X t (3.27) The low emissions sector supplies the remaining amount of energy (1 φ X ) to meet the energy demand. Since the stock of renewable input cannot be stored and hence has no inventories, the total output of the renewable energy simply equals the demand for renewable energy (3.28). y L t = (1 φ X )E f t (3.28) LC t = κ L y L t (3.29) UC L t = LC t y L t (3.30) The total cost of production LC t is given as the unit cost κ L times total output. This for example can include the cost of transforming renewables into energy. Since the low emissions energy sector does not hold inventories, the investment decisions are driven solely by the capital stock formation which is the same as the rms in the production sector. The price of energy, p E t is derived as follows: p E t = ( φ X UC X t (1 + τ X ) + (1 φ X )UC L t (1 + τ L ) ) (1 + θ) (3.31) or a simple weighted unit cost adjusted for industry specic taxes, τ X andτ L respectively times the xed mark up θ. Assuming κ L > κ X, which also hold empirically, then 3.31 implies that as the share of renewable energy in total energy supply goes up, the price of energy will increase as well. 3.3 Environment The environment is introduced in the model as providing the non-renewable resource X t through extraction from the ground and as absorbing Greenhouse Gasses (GHGs) in the atmosphere. While both the extraction process and the GHG accumulation process are subject to much controversial debate in literature (Nordhaus (1992); Tol (2002); Stern (2007); Weitzman (2009); Hope (2011); Pindyck (2013)) in this version of the model, both are introduced as simple linear functions. 11

12 The rate of depletion of the non-renewable source X t is endogenously given as: X t = X t 1 y X t (3.32) If the initial stock X 0 is very large and yt X comparatively small, then rate of depletion will be extremely small having a negligible extraction ination impact on the price of non-renewable energy (3.26). GHGs are assumed to accumulate at a linear rate relative to the level of rm production and of high emission energy sector production. The increase in stock is formalized as: GHG t = GHG t 1 (1 λ) + y t + y X t ξ Y G (3.33) where λ is an exogenously dened decay parameter, representing the absorption capacity of the environment and ξ Y G is the level of emissions added per unit of output produced. Innovation in emission eciency implies that ξ Y G will go up reducing the level of emissions per unit of output produced. Like the depletion rate of X, the accumulation rate of GHGs is dependent on the initial stock GHG 0 and the total output produceda high GHG 0 and/or a high growth rate of GHG t imply a more aggressive policy intervention to mitigate emissions. 3.4 Households Households comprise capitalists and employed or unemployed workers. In the model, all household agent categories are assumed to follow the same decision making procedures. The key dierence lies in the income source: Inc k t = Π f + Π X t + Π L t + Π b t + r d D k t 1 (3.34) Inc h t = W B t + r d D h t 1 (3.35) Inc u t = UB t (3.36) Capitalists earn prot income from rms, energy sector and commercial banks plus interest income on deposits they hold (3.34). Employed workers earn wages plus interest income from deposits (3.35), while the unemployment households receive transfers from the government (3.36). Given a total worker population of N, unemployed households are simply workers not employed by the rm sector (3.37). N u = N N f t (3.37) ub t = N u.ɛ (3.38) 12

13 The unemployed households N u are expected to maintain a socially dened minimum level of consumption ɛ in the form of unemployment benets ub t (3.38). The nominal value of the transfer program in is given as: UB t = ub t.p t (3.39) Household income after tax τ h gives the disposable income as follows: Y D t = Inc t (1 τ h ) (3.40) Households make consumption decisions based on real income and wealth levels. The consumption decision in real terms is dened as: c t = α 1 yd t 1 + α 2 v t 1 (3.41) where yd t and v t are real values of disposable income and wealth and α 1 is the marginal propensity to consume income and α 2 is the marginal propensity to consume wealth. Disposable income net of consumption results in a change in nominal wealth: Savings are allocated between cash M and deposits D. V t = Y D t C t (3.42) M t = ρc t (3.43) D t = V t M t (3.44) Households have a preference to keep a fraction ρ of consumption in the form of cash (3.43) and the remaining amount is deposited in commercial banks (3.44). 3.5 Commercial Banks Commercial banks in the model are kept relatively simple. against which loans are given out to the production sector. Holding deposits for households L b t = L f t + L X t + L R t (3.45) D b t = D k t + D h t (3.46) All loans as assumed to be provided on demand such that the total loans supplied equals L b t (3.45) against total household deposits D b t (3.46). 13

14 Since deposits are a liability against which the banks need to pay interest to households and give out loans, commercial banks have a preference to investment in treasury bills (TBills) issued by the government to ensure liquidity against any future changes in deposits. Thus if the commercial banks' net liquidity ratio (a measure of leveraging following Godley and Lavoie (2007, 11:337)) falls below a desired ratio ψ such that: NLR t = Db t L b t D b t < ψ (3.47) then the bank will purchase TBills to maintain the ratio such that the realized ratio equals: LR t = T B t + D b t L b t T B t + D b t (3.48) and the amount of TBills purchased can be derived as: [ L b T B t = Max t (1 ψ)dt b ], 0 (1 ψ) (3.49) The M ax condition implies that a bank cannot purchase negative TBills. If the commercial banks are unable to nance the purchase of TBills from their own deposits, then they take advances from the central bank which acts as a lender of last resort. The total value of advances is derived as: A b t = ( L b t D b t ) (1 ψ)zt (3.50) where z = 1 if NLR t < ψ or z = 0 otherwise implies that the total advances demanded are conditional upon two requirements; rst the current deposits must be insucient to allow for the procurement of desired level of TBills, and second, the NLR ratio should be less than the desired liquidity ratio of ψ (3.47). Bank prots are derived as Π b t = r l L b t 1 + r b B b t 1 r d D b t 1 r a A b t 1 (3.51) which is interest received on loans and TBills less interest paid on deposits and advances (3.51). As part of the borrowing and lending rates on interest norms, the interest rate on loans is kept higher than the interest rate on deposits such that r l r d. 3.6 Government The government plays two important roles in the model. First it ensures a certain consumption level dened exogenously as Γ which in nominal terms equals 14

15 G t = Γp t (3.52) and second it ensures a minimum consumption level for the unemployed such that the total unemployment benets bill is UB t (3.39). This expenditure is nanced through tax revenues that it earns from the rms and the households where total tax collected equals: T ax t = T f t + T X t + T R t + T k t + T h t (3.53) If the tax revenue is not sucient to nance the government expenditure then the government issues TBills. The debt requirement DR t of the government is thus dened as: DR t = G t + UB t + r b B t 1 T ax t Π CB t (3.54) where r b B t 1 is the interested owed on existing TBills issued and Π CB t redistributed to the government. The new TBills issued simply equal the debt requirement: are central bank prots B t = DR t (3.55) 3.7 Central Bank In the model, the central bank is assumed that acts as the nancial arm of the government rather than an independent regulator authority. The central bank issues advances to commercial banks on demand such that A CB t = A b t (3.56) and the endogenous money supply is based on household cash demand so total currency in circulation in the system equals: M CB t = M k t + M h t (3.57) The central bank is also assumed to purchase any TBills not held by the private sector: B CB t = B t B b t (3.58) Any prots earned by the central bank are fully redistributed to the government where prots equal: Π CB t = r b B CB t 1 + r a A CB t 1 (3.59) 15

16 4 Policy experiments This section introduces several policy experiments. The rst experiment shows a reduction in government expenditure. Following the literature on low or no-growth scenarios (Jackson (2009); Victor (2012); Victor and Jackson (2013)) the experiment shows that reducing demand will result in a reduction in output and income levels and subsequently emissions. The results discuss the adjustment process of the economy and the distributional eects resulting from higher unemployment levels. The second experiment shows a switch to a higher renewable energy production following some proponents of ecological economics literature (Berr (2009); Spash and Ryan (2012); Fontana and Sawyer (2013)). The third experiment highlights relative output and emissions trajectories given dierent levels of innovation. These experiments follow the decoupling literature that aim to establish relative growth rates between output productivity and energy production (Jackson (2009); Fischer-Kowalski, Swilling, von, RenY., Moriguchi, Crane, Krausmann, Eisenmenger, Giljum, Hennicke, Kemp, Romero, and Siriban-Manalang (2011)). The last set of experiments capture other strands of the literature by introducing an endogenous tax pegged to the level of emissions that negatively feed back on the capital stock (Rezai, Foley, and Taylor (2012); Pindyck (2013); Taylor and Foley (2014)). This shadow damage function approach allows for exploring possibilities where the economy can be forced to adjust its production process endogenous to the level of emissions. All simulations are compared to the benchmark business-as-usual (BAU) case calibrated on parameter values given in Appendix B. 4.1 Policy 1 - Reduction in government expenditure In the rst policy experiment, the exogenously dened government expenditure Γ is reduced by 10%. Figure 4.1 shows real output, income, and consumption levels of workers and capitalists. Figure 4.1: Policy 1 - Key indicators The reduction in government expenditure reduces total demand, which pushes down output as intended. In the short-run, due to lags in information updating, rms over-shoo their production capacity in the short-run, allowing output to fall below current demand. In order to meet the 16

17 demand-output shortfall, rms run down their inventories until output and sales equalize in the long-run at a lower equilibrium. Similarly, a reduction in demand reduces the real incomes of households. As consumption decisions adjust with a lag, households use their savings to smooth out demand in the short-run. Figure 4.2: Policy 1 - Unemployment As rms are the only employers in the model, a reduction in demand forces rms to lay o workers resulting in a higher level of unemployment which adds a dual burden on the government. First, tax revenue declines as output and income levels fall and second, unemployment transfers go up to ensure minimum income levels. 17

18 Figure 4.3: Policy 1 - Adjustment process Figure 4.3a shows the capital stock adjustment of the production sector driven by the target capacity utilization rate ν = 0.8. As a reduction in demand,leads rms to reduce their output and capital stock. Energy demand falls as well which induces the energy sector to adjust its capital stock. Figure 4.13a highlights the sensitivity of adjustment across the goods and the energy sectors. The energy sector's response shows a dampened response to a change in aggregate demand relative to rms. Lags in adjustment process play a key role in mitigating the shocks across sectors. Since rms are directly responding to changes in demand, they go through an over adjustment process. On the other hand, the energy sector is responding to a change in demand of the rms and hence adjusts its production process even more slowly. Based on model calibration the outcomes show that rms overshoot the adjustment process but achieve equilibrium faster which the energy sector over-adjusts by half but also takes twice the time to achieve equilibrium. Figure 4.3b shows the evolution of the prot share and the wage share (Bhaduri and Marglin (1990); Goodwin (1967); Taylor (2004)). As the reduction in demand reduces wages for workers, prots also start to fall implying a reduction in dividends for capitalists. As capitalists take time to update their expectations regarding prot income, rms push up the mark-up in the short-run increasing capitalist incomes and forcing the economy to grow through a prot led regime. This forces, the production to go up slightly resulting in lower unemployment levels, higher wage levels, and lower prot income expectations.thus the economy enters in a wage led regime in the long-run such that the prot share. 1 The full-adjustment process that shows a full Goodwin cycle results in the wage and prot stabilizing at the initial pre-shock levels of equilibrium. 1 In a Goodwin Cycle analysis, the direction of adjustment is reversed with a positive shock. The economy is wage-led in the short-run and prot led in the long-run. 18

19 4.2 Policy 2 - Higher renewable energy share In this experiment an exogenous shock increases the share of renewable energy in total energy supply from 5% to 10%. Figure 4.4: Policy 2 - Energy and Capital Intensity The change in the energy mix does not have a signicant impact on the long-run real output of the economy although the demand shifts signicantly from workers to capitalists increasing inequality in the system. This eect is not commonly discussed in the literature since distribributional eects are not a standard feature of many models. A key driver of this result are cost and pricing mechanisms that propagated through the rm and energy sectors decision making process that then exacerbate inequality. 19

20 Figure 4.5: Policy 2 - Energy Demand A higher share of renewable energy increases the price of energy (Figure 4.5e) and subsequently increases the input costs for rms (Figure 4.5a)thus raising the goods price. This dual price increase reduces the real incomes of workers, while increasing the prots for capitalists. As the overall demand for output falls slightly in the long-run, so does the energy demand. The nominal energy bill, however, is higher than the baseline value due to higher costs and prots in this sector. 20

21 Figure 4.6: Policy 2 - Capacity Utilization This energy policy also has capacity eects. Figure 4.6 shows the capacity adjustment process of the energy sector. Doubling the renewable energy production increases the capacity utilization to 100% where it remains until the sector stays at its peak in adapting the capital stock. Only then, does capacity utilization ratio in the energy sector decreases to the utilization ratio. 4.3 Policy 3- Increase in Capital and Energy Eciency In the BAU scenario a capital to output ratio ξ Y K = 1 and an energy to output ratio ξ Y E = 1 are assumed. In this section, both the capital stock and energy eciency is increased for both these parameters. That is, innovation boosts the output that can be produced with a given capital stock and energy input level. Table 4.1 summarizes three exogenous technological change scenarios that are tested using the model. Table 4.1: Policy 3 - Innovation scenarios Scenario ξ Y K ξ Y E Business as usual (BAU) BAU 1 1 Increase in capital eciency only I Increase in energy eciency only II Increase in capital and energy eciency III Scenario I increases capital eciency by 20% such that less capital is required to produce the same level of output. Similarly in scenario II, the energy eciency parameter sees a gain of 20% such that 20% less energy is required to produce the same level of output. In scenario III both the capital and energy eciency parameters increase by 20% each. 21

22 Figure 4.7: Policy 3 - Real macro balances Figure 4.7 summarizes the key results of simulations. The rst graph shows the adjustment process of real output; it stays roughly equal relative to the BAU in all the cases. On the other hand, the three policies have very dierent impacts on the household incomes. Scenario III, which is the most redistributive, shifts the demand from capitalists to workers, while scenario II, with pure energy eciency gains, has the smallest impact. Figure 4.8: Policy 3 - Prices and costs Figure 4.8 shows the evolution of prices in the system. As expected, unit costs decline for all scenarios with scenario III showing the highest reduction in per unit costs. On the other hand, even though scenarios I and II show the same unit costs, there is a signicant dierence in the markup. This subsequently results in relatively dierence prices for all three cases. These dierences are explained as followes: In scenarios I and III, a capital eciency gain results in a lower capital stock requirement, which reduces the burden of rms own investment subsequently allowing rms to ease othe markup. On the other hand, an energy eciency gain, lowers the input costs of energy forcing the unit cost to fall for scenario II. This in case of an eciency gain of both capital and energy to output (scenario III), both the unit costs fall, and the mark up falls reducing the price the most. 22

23 Figure 4.9: Policy 3 - Capital and Energy to Output ratio Figure 4.9 shows the capital and energy intensity of the economy relative to output. Since capacity utilization is set at 80%, actual physical capital is 25% higher than the output hence resulting in baseline value of For scenarios I and III, the capital-to-output ratio declines as a lower capital stock is required to produce the same level of output. Scenario II shows little in the capital to output ratio. The capital ecency gain in scenario I reduce the energy intensity of the economy as an indirect result of lower energy requirement for operating a lower level of capital stock. Figure 4.10: Policy 3 - Emissions 23

24 Figure 4.10 gives the level of unemployment and the level of emissions for three scenarios. While the level of unemployment on average remains close to the BAU value, scenario II reduces it slight more than the baseline value while scenario I increases it slightly. In the shortrun, the capital eciency driven adjustment policies create a higher volatility in the labor market, which raises unemployment levels signicantly. GHG accumulation dened through a simple linear function shows a slower rate of accumulation for all three scenarios. As expected, scenario III shows the strongest decline. 4.4 Policy 4 - Endogenous Tax This section explores various scenarios in which the tax rates are endogenized to the level of emissions. In order to fully track the impact of endogenous taxes on outputs, the experiments are conducted with a modied price mechanism where the tax rate is endogenized and a xed markup value θ = 0.1 is assumed (3.8). Assuming that the target steady state level of emissions is set exogenously at 380 ppmv (current level of GHG concentrations), the endogenous tax ination factor (T IF ) is dened as: T IF t = ( [ ]) GHGt Max, (4.1) The endogenous tax rate is determined as the base tax rate τ multiplied by the T IF t as follows: τ t = τ.t IF t (4.2) Thus if emissions deviate from the desired level of GHGs in the atmosphere, then the tax rate will increase. The Max condition in 4.1 implies that tax rates are not allowed to fall below the minimum tax level. Three taxation scenarios are dened in Table 4.2. Table 4.2: Policy 4 - Tax scenarios Scenario τ F τ H Business as usual (BAU) BAU Sales tax I 0.2.T IF t 0.2 Income tax II T IF t Sales and Income tax III 0.2.T IF t 0.2.T IF t In the model, two types of taxes are implemented, a sales tax imposed on rms (τ F ) and an income tax imposed on households (τ H ). For the purposes of simulations, both taxes are kept exactly equal to 20% in the BAU scenario. In scenario I, the sales tax is endogenized using the T IF over the baseline tax rate of 0.2. Similarly in scenario II, the income tax is endogenized using the T IF. Scenario III endogenizes both sales and income taxes. 24

25 Figure 4.11: Policy 4- Real macro balances Figure 4.11 shows overall macro balances for all three scenarios relative to the BAU scenarios. All taxation policies reduce real output, with the combined endogenous taxes (scenario III) showing the highest decline which is followed by the sales tax (scenario II). Income levels of both workers and capitalists decline as a result of all three taxation scenarios. While scenarios I and II have the same level of impact on the income of workers, capitalists see the lowest decline in income with just the income tax. Figure 4.12 shows the evolution of the cost, tax and the pricing structure in the system. As rms pass the burden of the tax on to the consumers, scenarios I and III show a signicant increase in unit costs.the endogenous income tax (II), as expected, does not impact rms cost decisions. On the other hand, Figure 4.12b shows that the endogenous income tax goes up by the highest amount relative to the other two tax scenarios. Figure 4.12: Policy 4 - Real macro balances 25

26 Figure 4.13: Policy 4 - Unemployment and emissions Figure 4.13 shows the change in unemployment levels and emissions. Tax scenario III leads to a signicant increase in the unemployment level as a result of the large fall in output. On the other hand, emissions also fall most in this case; they thus reach a much low steady stated compared to the BAU. 5 Conclusions This paper presents a highly stylized stock-ow consistent (SFC) closed economy macro model of a high income region that interlinks production with households, the energy sector, and the environment. Government and the nancial sectors are added to close the economic system. A key advantage of the SFC framework is that it allows for feed backs across all the key sectors of the economy which may result in unintended outcomes not captured in regular modeling frameworks. The aim of this model is to test various climate related policies discussed and debated in recent literature and discuss their impact on growth, unemployment, and distributions within the modeling framework. Four energy related policy experiments are conducted on the baseline model; reduction in government expenditure, higher share of renewable energy, relative gains in capital and energy eciency, and endogenous taxation pegged to the level of emissions. Each experiment traces the policy impact on the two production sectors in terms of output, prices, and employment levels. These then have an impact on the income levels of the two types of households; workers and capitalists which the unemployed households add to the burden of the government. Thus understanding distributional impact of policies are at the core of all climate related policies discussed in the paper. 26

27 A BAU calibration is used as a counter-factual scenario for all policy scenarios. Following the de-growth literature, the rst experiment shows the impact of a 10% reduction in government expenditure. While output and emissions go down, unemployment levels go up increasing the burden on the government which also has lower tax revenues due to a loss in output. The second experiment shows a forced higher share of renewable energy, which although reduces emissions, also results in higher inequality in the household sector. The third experiment shows the impact of exogenous capital and energy technological shocks which can result in lower inequality and lowering emissions. The last set of experiments highlight how demand can be endogenously controlled through various tax policies pegged to the level of emissions. This allows for endogenous demand side adjustments which then feedback into the production decisions of the rms and subsequently emissions. The next steps in the model development include calibrating the model the EU region based on the underlying patterns of production, consumption, energy mix, emissions and techonological eciency indicators. 27

28 References Berr, E. (2009): Keynes and Sustainable Development, International Journal of Political Economy, 38(3), Bhaduri, A., and S. Marglin (1990): Unemployment and the Real Wage: The Economic Basis for Contesting Political Ideologies, Cambridge Journal of Economics, 14(4), Caverzasi, E., and A. Godin (2013): Stock-Flow Consistent Model through the Ages, Levy Economics Institute Working Paper No. 7, January. European Commission (2014): European Economic Forecast: Winter 2014, Discussion paper, European Commission. Fischer-Kowalski, M., M. Swilling, W. E. von, RenY., Y. Moriguchi, W. Crane, F. Krausmann, N. Eisenmenger, S. Giljum, P. Hennicke, R. Kemp, L. P. Romero, and A. Siriban-Manalang (2011): UNEP(2011) Decoupling natural resource use and environmental impacts from economic growth. UNEP Regional Oce for Africa (ROA). Foley, D., L. Taylor, and A. Rezai (2013): The Social Cost of Carbon Emissions, INET Research Notes 28, Institute for New Economic Thinking (INET). Fontana, G., and M. Sawyer (2013): Post-Keynesian and Kaleckian thoughts on ecological macroeconomics, European Journal of Economics and Economic Policies: Intervention, 10(2), Godley, W., and M. Lavoie (2007): Monetary Economics: An Integrated Approach to Credit, Money, Income, Production and Wealth. Palgrave Macmillan, New York. Goodwin, R. M. (1967): A Growth Cyclepp Cambridge University Press, London. Hope, C. W. (2011): The social cost of CO2 from the PAGE09 model, Economics Discussion Papers , Kiel Institute for the World Economy. IPCC, I. P. o. C. C. (2007): Fourth Assessment Report: Climate Change Geneva: IPCC. (2012): Fifth Assessment Report: Climate Change Geneva: IPCC. Jackson, T. (2009): Prosperity without Growth: Economics for Finite Planet. Routledge. Kronenberg, T. (2010): Finding common ground between ecological economics and post- Keynesian economics, Ecological Economics, 69(7), Laing, T., M. Sato, M. Grubb, and C. Comberti (2013): Assessing the eectiveness of the EU Emissions Trading System, Grantham Research Institute on Climate Change and the Environment Working Papers 106, Grantham Research Institute on Climate Change and the Environment. Nordhaus, W. D. (1992): The 'DICE' Model: Background and Structure of a Dynamic Integrated Climate-Economy Model of the Economics of Global Warming, Cowles Foundation Discussion Papers 1009, Cowles Foundation for Research in Economics, Yale University. 28

29 Piketty, T. (2014): Capital in the Twenty-First Century. Cambridge Massachusetts: The Belknap Press of Harvard University Press. Pindyck, R. S. (2013): Climate Change Policy: What Do the Models Tell Us?, Journal of Economic Literature, 51(3), Rezai, A., D. Foley, and L. Taylor (2012): Global warming and economic externalities, Economic Theory, 49(2), Rezai, A., L. Taylor, and R. Mechler (2013): Ecological macroeconomics: An application to climate change, Ecological Economics, 85(C), Santos, C. H. D., and G. Zezza (2008): A Simplied, 'Benchmark', Stock-Flow Consistent Post-Keynesian Growth Model, Metroeconomica, 59(3), Spash, C. L. (2012): New foundations for ecological economics, Ecological Economics, 77(C), Spash, C. L., and A. Ryan (2012): Economic Schools of Thought on the Environment: Investigating Unity and Division, Cambridge Journal of Economics, 36(5), Stern, N. (2007): The Economics of Climate Change: The Stern Review. Cambridge University Press. Stockhammer, E. (2004): The Rise of Unemployment in Europe: A Keynesian Approach. Edward Elgar Publishing. Taylor, L. (2004): Reconstructing Macroeconomics: Structuralist Proposals and Critiques of the Mainstream. Harvard University Press. Taylor, L., and D. Foley (2014): Greenhouse Gasses and Cyclical Growth, INET Working Paper 38. Tol, R. (2002): Estimates of the Damage Costs of Climate Change. Part I: Benchmark Estimates, Environmental & Resource Economics, 21(1), Victor, P. (2008): Managing without Growth: Slower by Design, Not Disaster. Edward Elgar Publishing Ltd., Cheltenham, UK. Victor, P., and T. Jackson (2013): Developing a demographic sub-model and an Input- Output Structure for the Green Economy Macro-Model and Accounts (GEMMA), Discussion paper, Final Report CIGI-INET. Victor, P. A. (2012): Growth, degrowth and climate change: A scenario analysis, Ecological Economics, 84(C), Weitzman, M. L. (2009): On Modeling and Interpreting the Economics of Catastrophic Climate Change, The Review of Economics and Statistics, 91(1),

30 A Variables Variable Description A Advances B Treasury bills LR, LR N Liquidity Ratio (realized, notional) c, C Consumption (real, nominal) D Deposits DR Debt requirement E Energy EB, ED Energy bill, energy demand G Nominal government expenditure GHG Greenhouse Gasses i, I Capital investment (real, nominal) in, IN Inventories (real, nominal) Inc Income k, K Capital stock (real, nominal) L Loans M Money stock p Price Π Prots s, S Sales (real, nominal) T Tax u Capacity utilization rate ub, UB Unemployment benets (real, nominal) UC Unit cost v, V Wealth (real, nominal) W B Wage bill X Non renewable input y, Y Total rm output (real, nominal) yd, Y D Disposable income (real, nominal) 30

31 B BAU Parameter calibration Parameter Value Description N 2000 Workers N k 5% Capitalists as a % of total population Γ 800 Government expenditure ω 1 Wage rate per unit of labor α MPC out of income α MPC out of wealth β 0.2 Rate of closing gap between target and current capacity utilization δ 0.1 Rate of depreciation ν 0.8 Target capacity utilization κ X 1 Base extraction cost of non-renewable input X κ R 0.25 Base extraction cost of renewable input R τ 0.2 Tax rate σ 0.25 Target inventories to sales ratio γ 0.4 Rate of closing gap between target and current inventories θ 0.1 Markup on costs ρ 0.25 Demand for cash as a fraction of consumption ɛ 0.2 Minimum consumption ψ 0.05 Target liquidity preference ratio λ GHG decay rate ν 0.8 Target capacity utilization ratio r l 0.04 Interest on loans r d 0.02 Interest on deposits r b 0.02 Interest on TBills r a 0.02 Interest on advances ξ Y K 1 Output to capital stock ratio ξ KE 1 Capital stock to energy ratio ξ Y N 1 Output to labor ratio ξ Y X 1 Non-renewable input to high emissions capital ratio ξ Y L 1 Clean energy to clean energy capital ratio ξ Y G 10,000 Output to GHGs ratio 31