Do banks price environmental risk? Evidence from a quasi natural experiment in the People's Republic of China

Huang, Bihong; Punzi, Maria Teresa; Wu, Yu Working Paper Do banks price environmental risk? Evidence from a quasi natural experiment in the People's Republic of China ADBI Working Paper Series, No. 974 Provided in Cooperation with: Asian Development Bank Institute (ADBI), Tokyo Suggested Citation: Huang, Bihong; Punzi, Maria Teresa; Wu, Yu (2019) : Do banks price environmental risk? Evidence from a quasi natural experiment in the People's Republic of China, ADBI Working Paper Series, No. 974, Asian Development Bank Institute (ADBI), Tokyo This Version is available at: https://hdl.handle.net/10419/222741 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc-nd/3.0/igo/ ADBI Working Paper Series DO BANKS PRICE ENVIRONMENTAL RISK? EVIDENCE FROM A QUASI NATURAL EXPERIMENT IN THE PEOPLE’S REPUBLIC OF CHINA Bihong Huang, Maria Teresa Punzi, and Yu Wu No. 974 July 2019 Asian Development Bank Institute The Working Paper series is a continuation of the formerly named Discussion Paper series; the numbering of the papers continued without interruption or change. ADBI’s working papers reflect initial ideas on a topic and are posted online for discussion. Some working papers may develop into other forms of publication. Suggested citation: Huang, B., M.T. Punzi, and Y. Wu. 2019. Do Banks Price Environmental Risk? Evidence from a Quasi Natural Experiment in the People’s Republic of China. ADBI Working Paper 974. Tokyo: Asian Development Bank Institute. Available: https://www.adb.org/publications/do-banks-price-environmental-risk-evidence-quasi-naturalexperiment-prc Please contact the authors for information about this paper. Email: [email protected] Bihong Huang is a research fellow at the Asian Development Bank Institute. Maria Teresa Punzi is an assistant professor at Webster Vienna Private University. Yu Wu is an associate professor at Southwestern University of Finance and Economics Chengdu. The views expressed in this paper are the views of the authors and do not necessarily reflect the views or policies of ADBI, ADB, its Board of Directors, or the governments they represent. ADBI does not guarantee the accuracy of the data included in this paper and accepts no responsibility for any consequences of their use. Terminology used may not necessarily be consistent with ADB official terms. Working papers are subject to formal revision and correction before they are finalized and considered published. Asian Development Bank Institute Kasumigaseki Building, 8th Floor 3-2-5 Kasumigaseki, Chiyoda-ku Tokyo 100-6008, Japan Tel: +81-3-3593-5500 Fax: +81-3-3593-5571 URL: www.adbi.org E-mail: [email protected] © 2019 Asian Development Bank Institute ADBI Working Paper 974 Huang, Punzi, and Wu Abstract This paper maps the risk arising from the transition to a low-emission economy and studies its transmission channels within the financial system. The environmental dynamic stochastic general equilibrium (E-DSGE) model shows that tightening environmental regulations deteriorates firms' balance sheets as it internalizes the pollution costs, which consequentially accelerates the risks that the financial system faces. This empirical study, which employs the Clean Air Action that the Chinese government launched in 2013 as a quasi-experiment, supports the theoretical implications. The analysis of a unique dataset containing 1.3 million loans shows that the default rates of high-polluting firms rose by around 50% along their environmental policy exposure. At the same time, the loan spread charged to such firms increased by 5.5% thereafter, indicating that the banks do price the environment-related risks, but not sufficiently. Keywords: Environmental DSGE Model, Clean Air Action, lending spread, default rate JEL Classification: E32, E50, Q43, H23 “If some companies and industries fail to adjust to this new world, they will fail to exist.”1 1 Introduction The risks that environmental degradation and climate change pose for the stability of the financial system has been increasingly recognized (Bank of England (2019)). Physical risks arising from weather events such as more frequent or severe storms, floods and droughts dramatically increase the financial value at risk by destroying the assets like homes, offices and factories. Meanwhile, transition risks arising from changes in environmental policy and production technology requires a massive reallocation of capital and is likely to have significant and system-wide impacts on financial stability and might adversely a↵ect macroeconomic conditions (Network for Greening the Financial Systems (2019)). While the physical risks have been discussed for many years, transition risks are a relatively new category and remain underexplored. This paper fills this gap by investigating the mechanisms through which an environmental policy, aiming at improving air quality, a↵ects macroeconomy and financial sector. In particular, it sheds light on the channels through which the transition toward a low-emission economy transmit to a commercial bank’s balance sheet. As the regulations internalize the environmental costs, firms will face certain constraints on the cash flows and balance sheets, which will potentially weaken their solvency and consequently increase the risks facing the financial system. A growing number of countries are implementing a wide range of new rules and regulations that either tighten their existing environmental standards or impose entirely new environmental obligations on the business community. The long term benefits of such transition toward clean production are considerable, including cleaner air, improved health, reduced occurrence of natural disasters and sustainable economic growth.2. However, such policies can generate large transition costs for the economy, inducing fragility in the banking sector arising from compromised profitability for entrepreneurs, especially in the short run. Moderating the potential economic fluctuations arising from such policy change is hence of critical importance. However, most existing studies on the transition risks are partial and usually focus on the energy sector. This paper analyzes the economic and financial impacts of the transition towards low emission economy in the short term. We address this issue from both theoretical and empirical perspectives. We first develop an environmental dynamic stochastic general equilibrium (E-DSGE) model to illustrate how productive firms and financial institutions react to the imposition of environmental policy tools such as emission caps or intensity targets. We then employ a quasi-natural experiment to size the financial 1Open letter on climate-related financial risks, by Mark Carney (Governor of Bank of England) and Franois Villeroy de Galhau (Governor of Banque de France), available at https://www. bankofengland.co.uk/news/2019/april/open-letter-on-climate-related-financial-risks. 2See Fankhauser and Tol (2005), Mathiesen, Lund, and Karlsson (2011), Albrizio, Botta, Ko´zluk, and Zipperer (2014), Greenstone and Hanna (2014) and Kozluk and Zipperer (2015) 2 risks escalated by the adjustment process. In particular, we assess how the default rates and loan spread of firms with di↵erent level of emission intensity change following the Air Pollution Prevention and Control Action Plan (thereafter the Clean Air Action) that the Chinese government launched in 2013. This policy sets the national wide air pollution abatement target, with a special emphasis on the key regions of Beijing, Tianjin and Hebei (BTH), the Yangtze River Delta (YRD) and the Pearl River Delta (PRD).3This is the first time that the Chinese government has set quantitative air quality improvement goals with a clear time limit. Our theoretical model sheds light on a financial accelerator mechanism through which the prevention and mitigation of emission, in the form of tightening environmental policies, leads entrepreneurs to default endogenously on the outstanding value of their debt. This occurs because environmental policies a↵ect the value of collateral they use to pledge against borrowing, which in turn results in an excess premium on loans (i.e. the di↵erence between contracted loan rates and risk-free rates). The model extends the financial accelerator mechanism in the spirit of Kiyotaki and Moore (1997), Bernanke, Gertler, and Gilchrist (1999)andChristiano, Motto, and Rostagno (2014) by including an environmental policy that aims to reduce productive firms emissions. In particular, the financial accelerator mechanism assumes the existence of a dynamic feedback process between the value of capital and the level of borrowing, which can deplete when firms face tightening emission standards. The model assumes the existence of an optimal contract, which implies that entrepreneurs repay the debt to the bank if the value of their outstanding debt is lower than the value of the capital that they own. The tightening of environmental regulations on the use of polluting inputs in the production process reduces the asset values and expected return on projects. Consequently, entrepreneurs find themselves struggling and become unwilling or unable to perform their duties in line with the contractual conditions, as the repayment of debt is too costly relative to the rate of return. As the risk of default rises, the banking system will inevitably bear a financial loss due to its failure to obtain the expected earnings. Indeed, being aware of borrowers’ increasing exposure to environmental regulations, financial institutions, facing balance sheet distress, will charge higher lending rates as they internalize the environmental risks. The theoretical analysis of this research is related to Fischer and Springborn (2011) who incorporate environmental policy into a real business cycle model in which the total production is a function of capital, labor and polluting inputs. Within this framework, emissions are proportional to the use of polluting inputs and constrained by permit allocations so as to determine the maximum emission output ratio. We extend their model by assuming that a fixed cap on emissions constrains polluting inputs. This setting is consistent with the Chinese Clean Air Action that the government implemented in 2013 with the aim of improving the air quality in Chinese cities through varying 3Due to the short sample period (the Action Plan has been implemented only since 2013), we are unable to estimate or simulate the E-DSGE model. Therefore, the quasi-experimental analysis helps us to support the theoretical implications that a more restrictive climate change policy leads to environmental risk and loan defaults. 3 targets. The model assumes that higher production would generate more polluting emissions, which would then increase the shadow price of emission permits due to the fixed cap. A shock arising from a more severely tightened emission cap leads the output to fall as firms use fewer pollutant inputs, and the return on capital to decline as well. In contrast to Fischer and Springborn (2011), our model assumes that in each period entrepreneurs acquire new capital by borrowing new loans from financial intermediaries. The production function converts new capital into consumption goods, which households buy. The earned income and capital gains give entrepreneurs the capacity to repay their outstanding loans. However, a tightened constraint on emissions would reduce entrepreneurs’ profit, compromise firms liquidity, and endogenously raise the risk of default. In addition to the theoretical analysis, this research provides the clear identification and causal evidence by employing a unique micro-level dataset in a di↵erence-indi↵erence (DID) setting to investigate how an environmental policy shock a↵ects the default and lending spread, as well as their variations across di↵erent types of firms and banks. More specifically, we exploit the Clean Air Action that the Chinese government implemented in 2013 as a quasi-natural experiment. The Action sets quantitative air quality improvement goals for the entire country within a clear time frame. By 2017, the annual average concentration of coarse particulate matter (PM10) in urban area shall decrease by 10% compared with 2012, with the number of days with good air quality shall increase gradually. Moreover, it sets higher target for the key regions. During the same period of time, the annual average concentration of fine particulate matter (PM2.5) in Beijing-Tianjin-Hebei (BTH), the Yangtze River Delta (YRD), and the Pearl River Delta (PRD) shall decrease by 25%, 20%, and 15% respectively. To achieve these targets, the Action requires the local governments to tighten the regulations on heavily polluting industries. Given that Chinese firms largely rely on debt financing, the impact of environmental regulation can easily spill over to the banking sector. This naturally raises the question of whether banks consider the environmental shock when originating or extending credit to polluting firms. If banks are well-informed economic agents, they will in principle price in the increased default probability arising from the tightened environmental regulation. If not, they will underestimate an important source of risk for the sake of o↵ering more competitive loan rate. The issue of endogeneity hampers the inference of the causal impact of an environmental policy on financial stability. For example, some unknown factors, like economic conditions, may a↵ect the policy implementation and bank lending simultaneously. We employ several strategies to address the methodological challenge. First, we follow the recent literature, Jim´enez, Ongena, Peydr´o, and Saurina (2012)forinstance,byusing a unique micro-level dataset that contains 1.3 million loans that all types of banks granted to all non-financial firms located in six prefectures within Jiangsu province during the period of 2010 to 2016, moderating the concern that an analysis based only on macro data or bank-level data may su↵er from omitted-variables bias. Second, given that the emission intensity varies significantly across industries, we classify all the borrowers into high-polluting firms and low-polluting firms, and use the low-polluting 4 firms as a control group. Accordingly, the Clean Air Action allows us to implement a di↵erence-in-di↵erence (DID) estimation by using both before-and-after policy variation and cross-industry variation for identification. In other words, we compare the before-and-after change in the lending spread and default rates of firms with di↵erent pollution intensities. Moreover, considering that DID may not exclude time-varying variables that may bias the estimates, we further strengthen our identification by fielding loan and firm-time characteristics and several fixed e↵ects. For example, we include the important fixed e↵ects of bank*year to saturate the model with time-varying supplyside characteristics that might a↵ect the loan spread and city*year to control for yearly city-specific shocks. To understand whether the risks of lending to high polluting firms change following the policy enforcement, we trace the repayment status of loans granted during our sample period and find that the default risk of high polluting firms rose by around 50% after the policy implementation. However, compared with the period before the Action Plan, the loan spread for high-polluting firms increases by only around 5.5%. These finding represents salient evidence that the transition toward low emission economy has posed large risks on the financial stability. Although the banks are aware of such risks by raising the lending rate, but not to a degree comparable with the increased default rate. Further analysis indicates that the e↵ects of the Clean Air Action vary across firms of di↵erent sizes. Both the lending spread and the default rate show largest treatment e↵ects for the small and micro enterprises. Moreover, banks facing fewer government interventions are able to price the environmental risks more appropriately. This paper relates to di↵erent strands of the literature. First, the paper contributes to the literature that incorporates environmental factors into macroeconomy. Angelopoulos, Economides, and Philippopoulos (2010) analyze the impact of alternative environmental policy rules in a real business cycle model augmented with the assumptions that pollution occurs as a by-product of production process, and that only the government can engage in pollution abatement activity. Fischer and Springborn (2011) evaluate volatility and welfare costs by comparing cap-and-trade, the carbon tax, and the intensity target in a dynamic stochastic general equilibrium model with one polluting intermediate input. Heutel (2012) examines the optimal emission policy in a dynamic stochastic general equilibrium model with a pollution externality during phases of expansions or recessions. Annicchiarico and Di Dio (2015) analyze di↵erentenvironmental policy regimes in a new Keynesian model with nominal and real uncertainty and evaluate the transmission mechanism of shocks with the presence of nominal rigidities and a monetary authority. Tumen, Unalmis, Unalmis, and Unsal (2016)investigate the mechanisms through which environmental taxes on fossil fuel usage a↵ect the main macroeconomic variables in the short-run. Compared with the existing literature, this paper develops a financial accelerator mechanism that propagates the economic consequences of environmental risk by linking entrepreneurs to banks through the collateral value. Second, we contribute to the literature on endogenous default and financial accelerator. Christiano, Motto, and Rostagno (2014)developamodelwhereentrepreneurs 5 combine their own resources with loans to acquire raw capital, which can be converted into e↵ective capital in a process that is characterized by idiosyncratic uncertainty (i.e. risk shocks).4The authors prove that these risk shocks are important drivers of business cycle fluctuations. Similarly, Cesa-Bianchi and Fernandez-Corugedo (2018) study the e↵ect of macro uncertainty and micro uncertainty in a financial accelerator DSGE model with sticky prices. Forlati and Lambertini (2011), Quint and Rabanal (2014)andRabitsch and Punzi (2017) integrate risk shocks into the mortgage market, showing that default occurs endogenously when the housing investment risk increases. Spiganti, Comerford, et al. (2017) analyze the interaction between climate change policy and macroeconomic stabilization to exhibit the financial accelerator mechanism and fire sales of fossil fuel asset arising from carbon bubble. In particular, they assume that the government levies carbon taxes and provides green subsides to induce entrepreneurs to use zero carbon production technology, and find that such policies damage the balance sheets of entrepreneurs, with major macroeconomic implications due to the presence of financial frictions. Relative to those papers, the present paper incorporates the environmental risk into a DSGE models with endogenous default determined through a change in the balance sheets of entrepreneurs and the current asset holdings as collateral. Third, our empirical work enriches the growing literature studying the linkage of financial market to climate change and environmental risk. The potential e↵ect of such risk on financial stability is vigorously discussed by researchers and increasingly enters the agenda of regulators and supervisors (Carney (2015)). Trenberth, Dai, Van Der Schrier, Jones, Barichivich, Bri↵a, and Sheffield (2014) show that corporations’ production processes are vulnerable to natural disasters which is likely to be amplified by climate change. Bansal, Ochoa, and Kiku (2016) estimate the elasticity of equity prices to temperature fluctuations and find that global warming has a significantly negative e↵ect on asset valuations. Daniel, Litterman, and Wagner (2016)andGiglio, Maggiori, Stroebel, and Weber (2015) claim that stock and real estate market might help guide government policies if markets efficiently incorporate climate risks. De Grei↵, Delis, and Ongena (2018) suggest that the non-pricing of environmental policy exposure of fossil fuel firms leads to a carbon bubble. Hong, Li, and Xu (2019) show that prolonged drought in a country, measured by the Palmer Drought Severity Index (PDSI) from climate studies, forecasts both declines in profitability ratios and poor stock returns for food companies in that country. Dafermos, Nikolaidi, and Galanis (2018) report that climate change can have severe e↵ects on financial stability by increasing the rate of default resulting from lower firms’ profitability or lower asset prices, which they attribute to portfolio reallocation in the case of environmental damage. Despite the growing literature studying the direct impacts of pollution and climate change on financial assets, the research on the relationship between transition risks and financial stability are still scarce. This paper fills this gap by investigating the financial mechanisms through which the adjustment process toward a low-emission economy a↵ects macroeconomy and financial sector in the short term. 4Throughout the paper, I use the terms “risk shocks” and “uncertainty shocks” interchangeably. 6 2.4 Retailers The model assumes there is a continuum of retailers indexed n2[0,1] who transform intermediate goods Yt(n) into a final consumption good Yt, according to a constant elasticity of substitution technology: Yt=Z1 0 Yt(n)⇠1 ⇠dn ⇠ ⇠1 ,(2.13) Retailers aggregate intermediate goods from both green and non-green firms:14 Yt(n)=gYg e,t +ngYg e,t (2.14) where gand ng represents the market share of green and non-green firms, respectively. From standard profit maximization, input demand for the intermediate good iis obtained as: Yt(n)=✓Pt(n) Pt◆⇠ Yt,(2.15) where Pt(n)=gPg e,t(n)+ngPg e,t(n)andPtis the CES-based final (consumption) price index given by Pt=Z1 0 Pt(n)1⇠dn 1 1⇠ .(2.16) We assume a Calvo price-setting mechanism and retailers adjust each period their prices with a probability (1 ✓). P⇤ t(n) is the price that retailers are able to adjust. Thus, retailers maximize the following expected profit: max Et 1 X k=t (s✓)ktUCst+k UCst ⇢(P⇤ t(n) Pt+k Xt Xt+k )Y⇤ t+k(n) where Y⇤ t+k(i)=⇣P⇤ t(i) Pt+k⌘⇠Yt+k.X tis the markup of final over intermediate goods and in steady state is equal to X=⇠/(⇠1).The Calvo price evolves according to the following: Pt=h✓P⇠ t1+(1✓)(P⇤ t)(1⇠)i⇠ ⇠1.(2.17) Combining these two last equations, and after log-linearizing, we can obtain the following expression for the Phillips curve: ˆ⇡t=sEtˆ⇡t+1 ˆ Xt,(2.18) with =(1✓)(1)✓ ✓. 14Intermediate goods are perfect substitutes and this allows to have the same levels of intermediate goods prices according to whether they are produced by green or non-green firms. 13 2.5 Capital Producers Capital producers combine a fraction of the final goods purchased from retailers as investment goods, ik,t,to combine it with the existing capital stock, kt=Pjjkj e,t, in order to produce new capital. Existing capital is subject to an adjustment cost specified as k 2✓ik,t kt1 k◆2 kt1,where kgoverns the slope of the capital producers adjustment cost function. Capital producers choose the level of ik,t that maximizes their profits max ik,t qk tik,t  ik,t + k 2✓ik,t kt1 k◆2 kt1!. From profit maximization, it is possible to derive the supply of capital qk t=1+ k✓ik,t kt1 k◆, (2.19) where qk tis the relative price of capital. In the absence of investment adjustment costs, qk t,is constant and equal to one. The usual capital accumulation equation defines aggregate capital investment: ik,t =kt(1 k)kt1.(2.20) 2.6 Monetary Policy The Central Bank follows a Taylor-type rule that reacts to changes in inflation and output: Rt ¯ R=✓Rt1 ¯ R◆R⇣⇡t ¯⇡⌘⇡(1R)✓Yt y◆Y(1R) (2.21) where ⇡is the coefficient on inflation in the feedback rule, Yis the coefficient on output, and Rdetermines the degree of interest rate smoothing. 2.7 Market Clearing Yt=Ct+ik,t +Et+Mt+⌃ jµjGt+1(¯!j,t+1)qj,k t+1(1 k)kj e,t) (2.22) Ct=ct+gcg e,t +ngcng e,t +cb,t (2.23) kt=X j jkj e,t (2.24) 14 Lt=X j jLj t(2.25) bt=X j jbj e,t (2.26) qk t=X j jqj,k t(2.27) 2.8 Parameterization The time unit is measured in quarters. The parametrization follows standard values used in the real business cycle literature and they are reported in Table 2.15 The discount factor =bis set to 0.99 to target the annual nominal free-risk interest rate of 4%. Similar to Iacoviello (2015), entrepreneurs face a lower discount factor and e=0.94.The price elasticity ⇠is set equal to 6 and the Calvo probability to adjust prices, ✓, is set equal to 0.67. Similar to Justiniano, Primiceri, and Tambalotti (2015), the coefficient for the interest rate inertia, ⇢R, equal to 0.8, the reaction to the output gap, ⇢Y= 0.125, and the reaction to inflation of ⇢⇡= 1.5. The production function follows a constant returns to scale with a Cobb-Douglas specification, and the capital return to scale ↵is set equal to 0.35. The capital depreciation rate kis set at 0.025, while the adjustment cost parameter on investments is equal to 5. The share of clean energy and pollution emission in the production function, jis equal to 0.099 to imply an averaged energy expenditures as a share of GDP of 9.9%. The intensity target coefficient, #, is set equal to 0.05, a value smaller than j,asinFischer and Springborn (2011). In order to achieve a 20% decrease in the emissions, the persistence of the environmental policy shock is set equal to 0.97 and the standard deviation is set to 0.01. The size of green firms is set equal to 0.3. As in Christiano, Motto, and Rostagno (2014), the monitor cost is set to 0.21, and it is the same for both entrepreneurs, and the average probability of default, Fj(¯!j) is set to be equal to 0.007. Similar to Kollmann, Enders, and M¨uller (2011), the required bank capital ratio equal to 0.08, the bank cost parameter for deviating from capital requirements is set equal to 0.25. 3 Impulse Responses This Section presents results on simulated impulse responses under the scenario that the government enhances regulatory environmental standards to reduce the pollution emissions in the form of tightening the pollution constraint, i.e. ¯ Mshould decrease by 15We do not calibrate the idiosyncratic risk shock as we do not consider it. The variable !is introduced to allow for entrepreneurs’ defaulting behavior. 15 20%. The impulse responses show a percentage deviation from the initial steady state over a 20-quarter period under the environmental policy scenario. Fig. 1shows that such policy enforcement produces the e↵ect of reducing the productivity of non-green sector by around 2%, as firms have to use a lower input in the production process and the cap prevents additional output from being used as more of the intermediate good. Consequently, entrepreneurs in non-green sector decrease their investments and the price of capital drops down. Less investments leads to lower rental return on capital, as the demand for renting capital decreases. The decrease in the collateral value of non-green entrepreneurs, bring them to decrease their demand for funds. The change on the price of capital, capital stock and borrowing level a↵ects the entrepreneurs’ return on capital and the cut-o↵value, ¯ !ng, that endogenously determines the entrepreneurs’ failure to repay outstanding loans due to the rising costs of complying with environmental protection policies. Fig. 2shows that the cut-o↵value increases for both types of entrepreneurs, with higher impact on the non-green sector, as the regulatory environmental standards target exactly this sector in the economy (See solid line). This increase in the cut-o↵value reflects the movement in ¯!as described in Fig. 3. The left side corresponds to the distribution of default for the non-green sector, while the right side refers to the green sector. Fig. 3shows that when the governments tights the environmental standards in order to improve air quality, the default increases due a movement to the right of the ¯!,thus the default, measured by the shaded area, increases by the amount corresponding to the diagonal lines. However, the environmental policy shock leads to a change in the cut-o↵value also for the green sector. See dashed-dotted line in Fig. 2.Theincreasein the cut-o↵value is smaller relative to the non-green sector, reflecting a smaller movement to the right, as described in the right side of Fig. 3. Indeed, the default rate for the green sector is described by the diagonal lines area minus the shaded area. As a result, default rates in the non-green sector increase by around 0.75%. The existence of asymmetric information between bankers and entrepreneurs triggers banks to charge higher non-state contingent rates in face of expected higher monitoring costs, thus the excess premium, expressed as the di↵erence between contractual rates and risk-free rate, increases by around 0.5%. Through a banking capital channel, banks reduce the supply of loans as they face a reduction in bank capital due to higher default rates and due to a lower price of capital. As a result, banks deleverage because they keep in their balance sheet assets with lower value. This mechanism is reinforced by a banking funding channel, in which banks charge higher lending rates also to green entrepreneurs in order to recover from the losses from higher monitoring costs and the forgone loan repayments. Higher borrowing cost leads green entrepreneurs to lower their demand for external funds, and production and investment in the green sector slow down because of less financing available. Also clean energy inputs decrease as green entrepreneurs produce less. Ultimately, some green entrepreneurs can experience a lower return on new projects, a↵ecting their ability to repay their debt. Indeed, even if in smaller value, default rates in the green sector increase as well. This result is in line from the findings of the European Banking 16 Federation (EBF) which shows that Industrial and Commercial Bank of the PRC tend to have lower default rates to loans to green businesses relative to those in the non-green loans.16 To sum up, the main findings reveal that an environmental policy that aims at reducing pollution in the non-green sector spillovers to the green sector. 4 Background: Clean Air Action and the banking industry in the People’s Republic of China To provide further supporting evidences to the theoretical implications of the E-DSGE model, we employ the Clean Air Action that the Chinese government launched in 2013 as a quasi-natural experiment to examine the financial impacts of tightening environmental regulations in the short term. This section provides background information on the Clean Air Action and banking industry in the People’s Republic of China (the PRC). 4.1 Clean Air Action The main identification of this paper comes from the exogenous policy shock that the enforcement of Clean Air Action induced in 2013, which set the road map for air pollution control for the next five years in the PRC. Despite the phenomenal economic growth that the PRC achieved in recent decades, environmental degradation such as deteriorating water quality, land deforestation and pollution, frequent haze plague attracts a great deal of attention in the PRC. The year 2013 represents the start year of the PRC’s war on air pollution. On 1 January 2013, the Chinese government began publishing the air quality index (AQI), which measures fine particulate matter (PM2.5) per cubic meter, in real time in 74 cities throughout the country, making the worsening pollution quantifiable and visible to the public. Shortly thereafter, a massive fog and haze broke out in a fourth of the PRC’s territory, a↵ecting about 600 million people. In mid-January, Air Quality Index (AQI) in Beijing soared as high as 993, far exceeding the levels that the index defines as extremely dangerous. The population-weighted mean concentration of PM2.5 for the PRC as a whole was 54 µg/m3 in that year, with almost all the population living in areas exceeding the World Health Organization (WHO) Air Quality Guideline (Brauer, Freedman, Frostad, Van Donkelaar, Martin, Dentener, Dingenen, Estep, Amini, Apte, et al. (2016)). The haze with its unprecedentedly high index of PM2.5 concentration and extremely low visibility attracted global media attention and sparked outrage among the Chinese public, who eventually turning to be the “PM2.5 crisis.” Eight months after the widely-reported air pollution episode, on 12 September 2013, the the PRC’s State Council released the Action Plan for Air Pollution Prevention and 16See https://www.ebf.eu/wp-content/uploads/2017/09/Green-Finance-Report-digital. pdf. 17 Control.17 As a crucial step forward in fighting against air pollution, the Clean Air Action sets the road map for the next five years with a focus on three key regions - Beijing-Tianjin-Hebei (Jing-Jin-Ji), the Yangtze River Delta (YRD) and the Pearl River Delta (PRD). By 2017, for all the secondand third-tier cities, the annual average concentration of PM10 should decline by at least 10% compared with the 2012 level, and the number of days with clean air should increase. At the same time, the annual average concentration of PM2.5 should fall by 25%, 20%, and 15% respectively, for the three key regions. For Beijing, the annual average concentration of PM2.5 should remain at the 60 ug/m3 level. This is for the first time that the Chinese government had set quantitative air quality improvement goals for key regions with a clear time limit and key actions covering all the major aspects of air quality management. The new Air Pollution Prevention and Control Law that took e↵ect on January 1, 2016 later reinforced the Clean Air Action. It addresses pollution sources from coal, heavily polluting industries, vehicles, marine vessels and agricultural machinery, as well as the construction and food industries. Due to the urgency of severe air pollution, the stringency of the Action and the degree of its implementation are unprecedented (Sheehan and Sun (2014)). The main body of the Action specified the key targets, strategies, and measures, in many cases in the form of administrative orders from the government. After the nationwide Clean Air Action was announced, each provincial unit signed Letters of Responsibility with the Ministry of Environmental Protection and then issued its own version of Action by setting the reduction goals for annual average concentrations of PM10 or PM2.5. It sets clear target for the strategies and measures at the regional, sub-regional, sectoral, and sometimes firm levels. It usually divided the responsibilities for achieving the targets and implementing the measures e↵ectively among governmental departments. Manufacturing sectors were among the foci of the Action. Industrial upgrading and restructuring are necessary for high polluting industries with high energy consumption and with backward productivity or excess capacity. The strategies and measures targeting manufacturing sectors mainly include end-of-pipe measures, optimizing the industrial structure, promoting cleaner production and eco-industrial parks, and adjusting the structure of the energy supply and consumption. The application and upgrading of the removal technologies of SO2,NOx, particulate matters, and volatile organic compounds (VOCs) in key polluting sectors will be mandatory. The emission intensity in key industries should reduce by over 30%. Outdated production lines and small polluting firms should close. The entry requirements of highly polluting and energy-consuming sectors, such as iron and steel, cement, electrolytic aluminum, and coking, require strengthening, and the formulation of ban lists for the construction and expansion of industrial projects in these sectors is necessary. It should be compulsory to carry out environmental impact assessment (EIA) and energy-saving examination before the construction, transformation, and expansion of industrial projects. The approval of an EIA should take into account the total emissions of SO2,NOx, particulate matters, and VOCs as prerequisites. Banks should not be allowed to provide loans to 17For the details of the plan, please refer to http://www.cleanairthePRC.org/product/6349.html 18 projects that have failed to pass an EIA and energy saving examinations. The Action forbids regions and industrial sectors that have failed to achieve air pollution reduction goals from building new projects that would emit the same nonattainment pollutant. It also sets targets in terms of the structure of energy consumption to reduce coal consumption and promote renewable energy sources. It provides annual implementation plans under the multi-year plans. It specifies the targets, measures, and projects that require completion within each year. It expects the governmental departments to seek policy, funding, and technological support from corresponding ministries or departments of higher-level governments. It is also possible to establish special plans targeting major polluting sectors. 4.2 Banking Industry in the PRC Banks play a very important role in the Chinese economy as most Chinese firms largely rely on debt financing. The total assets of the Chinese banking system amounts to 268.2 trillion yuan (or US$38.9 trillion) by the end of 2018. It is roughly 3 times the size of the countries annual GDP and overtakes the eurozone’s banking assets. Before 1978, the banking system in the PRC was a mono-bank system. A single bank, Peoples bank of the PRC (PBoC), functioned both as a central bank and as a commercial bank, in charge of all businesses such as deposits, lending, foreign exchange, and monetary policy. As part of the economic reform, the financial system has become more diversified since 1978. The establishment of four state-owned specialized banks in 1983 aimed to take charge of commercial businesses. The Industrial and Commercial Bank of the PRC (ICBC) focused on the corporate lending, the Agriculture Bank of the PRC (ABC) aimed to promote the economic development in the rural areas, the Bank of the PRC (BOC) specialized in the foreign exchange business, and the the PRC Construction Bank (CCB) was responsible for construction and infrastructure developments. At the same time, the mandate of PBoC was the role of a central bank. In addition to these four state-owned specialized banks, various types of financial institutions started to emerge in the late 1980s. Established in 1987, the Bank of Communications (BoCom) was the first joint equity banks in the PRC. Although BoCom is technically a joint equity bank, it is more or less the same as the Big Four in terms of the regulation and political hierarchy. Both the four state-owned banks and BoCom are under the direct control of the central government and are held by the Ministry of Finance and a sovereign wealth fund the the PRC Investment Corporation. These Big Five belong to the top tier of the PRCs banking system, controlling for approximately 45% of the market share. The second tier contains the 12 joint equity commercial banks (JECBs), which are also mainly state-owned, while they have far fewer branches than the big five banks and banks operate their businesses relatively locally. The rest of the financial institutions such as rural credit cooperatives, city commercial banks, trust and investment cooperation, finance company, foreign banks, belong to the third tier. After the entry of the WTO in 2001, the Chinese financial system experienced several further reforms. In 2003, the government established the Chinese banking reg19 ulatory commission (CBRC) to monitor commercial bank operations. To improve the corporate governance of banks, it allowed the four state-owned banks to go to public from 2005 to 2010, and encourage city commercial banks to bring in foreign strategic investors, go public, reconstruct, and operate across regions. In 2006, the government completely opened the RMB business to foreign banks. The entry of foreign banks improves the efficiency of the Chinese banking system (Xu, 2011). As a result of reform, the proportion of assets of state-owned commercial banks decreased from 58.03% in 2003 to 37.29% in 2016, while the assets of joint-stock commercial banks increased from 10.70% in 2003 to 18.72% in 2016. According to the newest statistics released by the PRC Banking and Insurance Regulatory Commission (CBIRC), which replaced the CBRC in April 2018, there are 4588 financial institutes by the end of 2018, including 134 city commercial banks, 1427 rural commercial banks, 1616 village banks, 812 rural credit cooperative, 115 foreign banks, among others.18 5 Data and Empirical Strategy 5.1 Data source and summary statistics Our data come from a corporate credit database that the CBIRC Jiangsu Office established. With a population of 80.4 million in 2018 and an area of 102,600 km2, Jiangsu is one of the most densely populated provinces in the PRC. Thanks to its large and well-developed manufacturing sector, it is one of the PRCs fastest developing provinces over recent decades. As of 2018, Jiangsu had a GDP of US$1.377 trillion (RMB9.2 trillion), the second highest in the PRC (just after Guangdong), but greater than those of Mexico and Indonesia. However, with an economic structure in which the secondary industry accounts for around 40% of GDP and home to many of the world’s leading exporters of electronic equipment, chemicals and textiles, Jiangsu faces serious environmental degradation. In 2013, the industrial SO2 and COD emissions per unit of land area are 8.48 and 19.51 tons per square kilometer, respectively.19 This dataset contains around 1.3 million commercial loans that all banks operating in six prefectures within this province granted to all non-financial firms during the period of 2010 to 2016, allowing us to identify the causal e↵ect of environmental policy on the stability of the financial system by exploiting the variations across prefectures, banks, industries and borrowing firms. Since the database includes all loans that the banks granted within the jurisdiction, we eliminate concerns about sample selection. Moreover, this coastal province o↵ers an ideal setting in which to investigate how the process of adjustment toward low emission economy a↵ects the financial risks because it has a diverse economy with various types of banks. On one hand, the GDP per capita of the six prefectures varies widely between USD 7,000 and 20,000, representing di↵erent levels of economic development. On the other hand, various types of banks such as the 18The figures used in the paragraph come from the PRC Banking Regulatory Commission Annual Reports published in various years. 19The figures are calculated from the PRC Environmental Yearbook. 20 Big Five state-owned commercial banks, joint-equity commercial banks, foreign banks, city commercial banks, rural commercial banks, rural credit cooperative,. According to the newest statistics released by CBIRC, the total asset of commercial banks in Jiangsu Province amounts to RMB 16781.52 billion yuan (around US$2,494 billion), accounting for 8% of the whole commercial banking industry in the country as of 2018.20 The number of borrowers in this dataset amounts to around 100,000 firms, covering all industrial sectors in accordance with the classification defined by the Chinese government. This information allows us to identify the borrowers belonging to the highly polluting industries that the Clean Air Action targets. Besides the comprehensive coverage, the dataset provides detailed loan-level information, specifically a unique firm identifier, firm-level fundamentals (e.g., age, size, ownership and location), banks’ information (e.g., the ownership, the names and location of branches), and loan-level characteristics (e.g., loan amount, loan maturity, credit guarantee, issuing date, maturity date, and loan delinquency status). The banks update the loan information mandatorily with a monthly frequency throughout its whole life cycle. In this way, we can trace the repayment of loans and determine whether the banks properly price the risk of default, which the environmental regulation might escalate. In addition to default, loan spread is one of our main outcome variables. Following the existing literature, we measure the spread of each loan by the percentage deviation of its lending rate from the benchmark rate. This calculation allows us to rule out the change of the credit cost arising from adjustment of benchmark interest rate. Given that the commercial loans granted by the domestic commercial banks shall reflect the market response to the environmental risk in a better way, we remove all the loans granted by the development bank, policy banks and foreign banks from our analysis. Table 3provides the summary statistics of our main variables for the period of 1 January 2013 to 31 December 2014. Overall, the 52 commercial banks, including Big Five, 12 joint equity commercial banks, 1 postal saving bank, 8 city commercial banks, and 27 rural commercial banks, granted 379,130 loans to 59,094 firms during this period of time. The mean borrowing rate is 7.4%, 23.5% higher than the benchmark rate of 6%. The average amount of borrowing is RMB8.06 million. In terms of maturity, 93.6% of loans are short term borrowing. There are various types of loans, among which 45% are secured loans with collateral and around 40% are loans with a guarantee. With an average age of 10.6 years, 84.5% of borrowers are micro and small firms, 12.2% are medium-sized firms and 3.3% are big firms.21 The Big Five and rural commercial banks are the major lenders, accounting for 34.3% and 38.5% of loans respectively. 20The figures are calculated from the statistics released on websites of CBIRC (http://www.cbrc.gov.cn/chinese/home/docView/ C990691733D644B39582DEFA3EF1EF69.html) and CBIRC Jiangsu Office (yrlhttp://www.cbrc.gov.cn/jiangsu/docPcjgView/8AB96DDF7DDF487C95D1D4D4FB8FC0E1/600811.html) 21A firms size is defined as small and micro, medium, or large, based on The Standards of SMEs jointly issued by the PRCs Ministry of Industry and Information Technology, National Bureau of Statistics, National Development and Reform Commission, and Ministry of Finance. 21 5.2 Empirical Strategy This paper employs the Clean Air Action that Jiangsu province implemented in January 2014 as a quasi-experiment to evaluate the financial risks posed by the transition toward low emission economy. In particular, we rely on the DID approach to infer the impact of tightened environmental regulation on default and lending spread of bank loans. DID analysis consists of comparing the pre-post di↵erence in an outcome variable between a treatment and a control group. Specifically, for each loan, we classify the borrowers belonging to the highly polluting industry defined by the Clean Air Action (Jiangsu version) as our treatment group,22 while the rest as the control group. This approach has an advantage over simply comparing the outcome before and after the regulatory shock because there might be before-after di↵erences in the outcome that are due to broader trends. This is why having a comparison group, which is unexposed (or less exposed) to the policy shock, allows us to capture this trend and thus better estimate a counterfactual. Table 4compares the descriptive statistics between the control and treatment groups. The 52 commercial banks grant 294, 664 loans to the low polluting firms and 84,466 loans to the highly polluting firms respectively. The lending cost is similar across two groups while the default rate of highly polluting firms is slightly higher than that of low polluting firms. However, the average loan amount borrowed by the low polluting firms is twice as large as that by highly polluting firms. The term of maturity and loan type are also similar between the treatment and control groups. Regarding the firm size, big firms account for 3.7% of low polluting borrowers and 2% of high polluting borrowers. In terms of lenders’ structure, local banks including the city commercial banks and rural commercial banks grant 57.9% of loans to highly polluting firms while only 50.7% of loans to low polluting firms. Overall, we find that the treatment and control groups are comparable in loan characteristics. Within this framework, we implement the DID analysis to compare the default rate and loan spread of the high-polluting firms that the Clean Air Action specially targets with those of low-polluting firms with less exposure to the regulation. If the financial institutions like banks are aware of the environmental transition risks, their lending decisions regarding the high-polluting firms should di↵er from those regarding the other firms. This comparison, considers the loans granted to the high-polluting firms as the treatment group and the loans granted to the low-polluting firms as the control group. We obtain our DID estimators measuring the e↵ect of the environmental policy shock on the financial stability using the following model: ylbft =0+1Actiont+2Actiont⇤Treati+3Treatedi +4Lit +5Fft +ulbft (5.1) 22The high polluting industries targeting by the Clean Air Action (Jiangsu version) include steel, cement, thermal power, textile, chemical, petrochemical, nonferrous metal melting, sintering pellet, ferroalloy, steel rolling, coking, coating and plating, pharmaceutical, plastic, furniture, building materials, automotive repair and maintenance. 22 di↵erence-in-di↵erence estimation indicates that following the policy implementation the default rates of lending to the high-polluting firms that the Action targets dramatically increase by 50%. At the same time, loan spreads of these lending also rise, but to a much smaller degree, indicating that the commercial banks have not sufficiently priced the transition risk. Our empirical evidences are consistent with the theoretical implications of the environmental dynamic general equilibrium (E-DSGE) model which predicts higher default and lending rates when the model includes environmental policy shift such as the implementation of emission cap. The solid findings of this research suggests that transition towards low emission economy is one source of structural change which significantly a↵ects all economic sectors and the financial stability. While urgent action is desirable for environmental improvement, an orderly and smooth transition providing adequate time for technological progress could minimize these risks. In addition, financial institutions shall be aware of potential risks arising from the environmental adjustment process and embed them in their risk management and pricing strategies. Given that maintaining financial stability is within the mandates of central banks and financial regulators, it is necessary for them to integrate the monitoring of environment and climate-related financial risks into the prudential supervision to ensure the resilience of the financial system to the potential risks. 29 References Albrizio, S., E. Botta, T. Ko´ zluk, and V. Zipperer (2014): “Do environmental policies matter for productivity growth?,” . Angelopoulos, K., G. Economides, and A. Philippopoulos (2010): “What is the best environmental policy? Taxes, permits and rules under economic and environmental uncertainty,” . Annicchiarico, B., and F. Di Dio (2015): “Environmental policy and macroeconomic dynamics in a new Keynesian model,” Journal of Environmental Economics and Management, 69, 1–21. Bank of England, B. (2019): “Climate change: what are the risks to financial stability?,” Discussion paper, https://www.bankofengland.co.uk/knowledgebank/ climate-change-what-are-the-risks-to-financial-stability. Bansal, R., M. Ochoa, and D. Kiku (2016): “Climate change and growth risks,” Discussion paper, National Bureau of Economic Research. Becker, R., and V. Henderson (2000): “E↵ects of air quality regulations on polluting industries,” Journal of political Economy, 108(2), 379–421. Berman, E., and L. T. Bui (2001a): “Environmental regulation and productivity: evidence from oil refineries,” Review of Economics and Statistics, 83(3), 498–510. (2001b): “Environmental regulation and labor demand: Evidence from the south coast air basin,” Journal of Public Economics, 79(2), 265–295. Bernanke, B. S., M. Gertler, and S. Gilchrist (1999): “The financial accelerator in a quantitative business cycle framework,” Handbook of macroeconomics,1, 1341–1393. Brauer, M., G. Freedman, J. Frostad, A. Van Donkelaar, R. V. Martin, F. Dentener, R. v. Dingenen, K. Estep, H. Amini, J. S. Apte, et al. (2016): “Ambient air pollution exposure estimation for the global burden of disease 2013,” Environmental science & technology, 50(1), 79–88. Carney, M. (2015): “Breaking the Tragedy of the Horizon–climate change and financial stability,” Speech given at Lloyds of London, 29, 220–230. Cesa-Bianchi, A., and E. Fernandez-Corugedo (2018): “Uncertainty, Financial Frictions, and Nominal Rigidities: A Quantitative Investigation,” Journal of Money, Credit and Banking, 50(4), 603–636. Christiano, L. J., R. Motto, and M. Rostagno (2014): “Risk shocks,” American Economic Review, 104(1), 27–65. Cohn, J., and T. Deryugina (2018): “Firm-Level Financial Resources and Environmental Spills,” Discussion paper, National Bureau of Economic Research. Dafermos, Y., M. Nikolaidi, and G. Galanis (2018): “Climate change, financial stability and monetary policy,” Ecological Economics, 152, 219–234. Daniel, K. D., R. B. Litterman, and G. Wagner (2016): “Applying asset pricing theory to calibrate the price of climate risk,” Discussion paper, National Bureau of Economic Research. De Greiff, K., M. Delis, and S. Ongena (2018): “Being Stranded on the Carbon 30 Bubble,” Climate Policy Risk and the Pricing of Bank Loans. CEPR Discussion Papers. Fankhauser, S., and R. S. Tol (2005): “On climate change and economic growth,” Resource and Energy Economics, 27(1), 1–17. Fischer, C., and M. Springborn (2011): “Emissions targets and the real business cycle: Intensity targets versus caps or taxes,” Journal of Environmental Economics and Management, 62(3), 352–366. Forlati, C., and L. Lambertini (2011): “Risky mortgages in a DSGE model,” International Journal of Central Banking, 7(1), 285–335. Giglio, S., M. Maggiori, J. Stroebel, and A. Weber (2015): “Climate change and long-run discount rates: Evidence from real estate,” Discussion paper, National Bureau of Economic Research. Greenstone, M. (2002): “The impacts of environmental regulations on industrial activity: Evidence from the 1970 and 1977 clean air act amendments and the census of manufactures,” Journal of political economy, 110(6), 1175–1219. Greenstone, M., and R. Hanna (2014): “Environmental regulations, air and water pollution, and infant mortality in India,” American Economic Review, 104(10), 3038–72. Greenstone, M., J. A. List, and C. Syverson (2012): “The e↵ects of environmental regulation on the competitiveness of US manufacturing,” Discussion paper, National Bureau of Economic Research. Heutel, G. (2012): “How should environmental policy respond to business cycles? Optimal policy under persistent productivity shocks,” Review of Economic Dynamics, 15(2), 244–264. Hong, H., F. W. Li, and J. Xu (2019): “Climate risks and market efficiency,” Journal of Econometrics, 208(1), 265–281. Huang, B, Z. Y., and M. Zhou (2019): “Bottom up or Bottom down? Strategic environmental regulation and inbound foreign direct investment in China,” Discussion paper, ADBI Working Paper. Iacoviello, M. (2015): “Financial business cycles,” Review of Economic Dynamics, 18(1), 140–163. Jaffe, A. B., S. R. Peterson, P. R. Portney, R. N. Stavins, et al. (1995): “Environmental regulation and the competitiveness of US manufacturing: what does the evidence tell us?,” Journal of Economic literature, 33(1), 132–163. Jim´ enez, G., S. Ongena, J.-L. Peydr´ o, and J. Saurina (2012): “Credit supply and monetary policy: Identifying the bank balance-sheet channel with loan applications,” American Economic Review, 102(5), 2301–26. Justiniano, A., G. E. Primiceri, and A. Tambalotti (2015): “Household leveraging and deleveraging,” Review of Economic Dynamics, 18(1), 3–20. Kim, T., and Q. Xu (2017): “Financial constraints and corporate environmental policies,” Discussion paper, HKUST Finance Symposium. Kiyotaki, N., and J. Moore (1997): “Credit cycles,” Journal of political economy, 105(2), 211–248. 31 Kollmann, R. (2013): “Global Banks, Financial Shocks, and International Business Cycles: Evidence from an Estimated Model,” Journal of Money, Credit and Banking, 45(s2), 159–195. Kollmann, R., Z. Enders, and G. J. M¨ uller (2011): “Global banking and international business cycles,” European Economic Review, 55(3), 407–426. Kozluk, T., and V. Zipperer (2015): “Environmental policies and productivity growth,” OECD Journal: Economic Studies, 2014(1), 155–185. Kr¨ uger, P. (2015): “Climate change and firm valuation: Evidence from a quasinatural experiment,” Swiss Finance Institute Research Paper, (15-40). Mathiesen, B. V., H. Lund, and K. Karlsson (2011): “100% Renewable energy systems, climate mitigation and economic growth,” Applied energy, 88(2), 488–501. Morgenstern, R. D., W. A. Pizer, and J.-S. Shih (2002): “Jobs Versus the Environment: An Industry-Level Perspective,” Journal of Environmental Economics and Management, 43(3), 412 – 436. Network for Greening the Financial Systems, N. (2019): “A call for action: climate change as a source of financial risk,” Discussion paper, https://www.banque-france.fr/sites/default/files/media/2019/04/ 17/ngfs_first_comprehensive_report_-_17042019_0.pdf. Quint, D., and P. Rabanal (2014): “Monetary and macroprudential policy in an estimated DSGE model of the euro area,” International Journal of Central Banking, 10(2), 169–236. Rabitsch, K., and M. T. Punzi (2017): “Borrower heterogeneity within a risky mortgage-lending market,” . Sheehan, Peter, C. E. E. A., and F. Sun (2014): “China’s response to the air pollution shock,” Nature Climate Change, 4, 306–309. Spiganti, A., D. Comerford, et al. (2017): “The Carbon Bubble: Climate Policy in a Fire-sale Model of Deleveraging,” in 2017 Meeting Papers,no.734.Societyfor Economic Dynamics. Tanaka, S., W. Yin, and G. H. Jefferson (2014): “Environmental regulation and industrial performance: Evidence from China,” Work. Pap., Tufts Univ., Medford, MA. Trenberth, K. E., A. Dai, G. Van Der Schrier, P. D. Jones, J. Barichivich, K. R. Briffa, and J. Sheffield (2014): “Global warming and changes in drought,” Nature Climate Change, 4(1), 17. Tumen, S., D. Unalmis, I. Unalmis, and D. F. Unsal (2016): “Taxing fossil fuels under speculative storage,” Energy Economics, 53, 64–75. Walker, W. R. (2011): “Environmental regulation and labor reallocation: Evidence from the Clean Air Act,” American Economic Review, 101(3), 442–47. (2013): “The transitional costs of sectoral reallocation: Evidence from the clean air act and the workforce,” The Quarterly journal of economics, 128(4), 1787– 1835. 32 Fig. 1: Environmental Shock to Decrease Pollution Emissions 33 Fig. 2: Change in ¯!due to Tightness of Environmental Standards Fig. 3: Distribution of Default (F(!i j)) and Tightness of Environmental Standards: Non-Green Sector (Left side); Green Sector (Right side). 0.0 0.5 1.0 1.5 2.0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.5 1.0 1.5 2.0 0.0 0.2 0.4 0.6 0.8 1.0 34 Fig. 4: The dynamic impact of the Clean Air Action on default. The figure plots the impact of the Clean Air Action on default. We consider a 19-month window, spanning from 7 months before the Clean Air Action was implemented until 12 months after it was enforced. We control year, bank and prefecture fixed e↵ect. 35 Table 1: Bank Balance Sheet Assets Liabilities Green Firms (bg e,t+1)Domestic Deposits (dt) Non-Green Firms (bng e,t+1)Bank Capital (xt) Table 2: Parameters’ Values Parameter Description Value Households Discount factor 0.99 eEntrepreneurs Discount factor 0.94 cElasticity of substitution for consumption 1.01 ⌫LLabor disutility parameter 1 ⌘Labor supply aversion 1.01 kCapital depreciation parameter 0.025 kCapital adjustment cost 5 ↵Capital Share 0.35 gEnergy Inputs Share 0.099 ng Pollution Inputs Share 0.09 #Intensity Target coefficient 0.05 Fj(¯!j)Probability of default 0.007 ⇠Price Elasticity of Demand for Good n6 ✓Calvo’s Price Parameter for Nominal Rigidities 0.67 ⇢RMonetary Policy Inertia 0.8 ⇢YMonetary Policy Reaction to Y0.125 ⇢⇡Monetary Policy Reaction to ⇡1.005 bBanks Discount factor 0.99 ⇢bBanks Capital ratio 0.08 ⇥Cost of deviation from the required capital ratio 0.25 µjMonitoring Cost 0.21 gSize of Green Firms 0.3 ⇢MPersistency of Environ. Policy shock 0.97 MStandard deviation on Environ. Policy shock 0.01 36 Table 3: Summary Statistics This table presents the summary statistics of the key variables for the sample period running from Jan 2013 to Dec 2014. We report the summary statistics for the main outcome variable the lending spread – the percentage deviation of its lending rate from the benchmark rate; the loan-level characteristics including loan amount, maturity, and credit guarantee; the firm-level fundamentals of age and size; types and ownership of banks; and local economic structure and GDP per capita. Variable Obs Mean Std. Dev. Min Max Lending spread 379130 0.235 0.269 -0.4 2 Lending rate 379130 0.074 0.016 0.034 0.191 Benchmark interest rate 379130 0.060 0.227 5.6 6.55 Loan amount (CNY 10 thousand) 379130 806 2440 5 210000 Maturity Short term loan 379130 0.936 0.245 0 1 Mid-term loan 379130 0.05 0.217 0 1 Long-term loan 379130 0.015 0.123 0 1 Loan type Secured loan 379130 0.446 0.497 0 1 Fiduciary loan 379130 0.028 0.164 0 1 Loan on guarantee 379130 0.397 0.489 0 1 Pledged loan 379130 0.078 0.268 0 1 Discount loan 379130 0.052 0.222 0 1 Firm size Micro and Small firms 379130 0.845 0.362 0 1 Medium firms 379130 0.122 0.327 0 1 Big firms 379130 0.033 0.179 0 1 Company age (Year) 379130 10.6 5.73 1 60 Bank type Big five 379130 0.343 0.475 0 1 Joint-stock commercial banks 379130 0.134 0.341 0 1 City commercial banks 379130 0.138 0.345 0 1 Rural banks 379130 0.385 0.487 0 1 Local Economic structure Share of secondary industry 379130 0.505 0.018 0.442 0.526 Share of tertiary industry 379130 0.455 0.028 0.384 0.484 GDP per capita (CNT Yuan) 379130 101248.9 29247.46 35484 129926 37 Table 4: Summary statistics, low polluting versus highly polluting Industries This table compares the summary statistics of the key variables between low polluting and highly polluting industries for the sample period between Jan 2013 and Dec 2014. We report the summary statistics for the main outcome variable the lending spread – the percentage deviation of its lending rate from the benchmark rate; the loan-level characteristics including loan amount, maturity, and credit guarantee; the firm-level fundamentals of age and size; types and ownership of banks; and local economic structure and GDP per capita. Low polluting industry Highly polluting industry Std. Std. Variable Obs Mean Dev. Min Max Obs Mean Dev. Min Max Lending spread 294664 0.232 0.276 -0.4 2 84466 0.244 0.242 -0.4 2 Lending rate 294664 0.074 0.017 0.034 0.191 84466 0.074 0.015 0.034 0.18 Benchmark int. rate 294664 0.06 0.002 0.056 0.0655 84466 0.06 0.002 0.056 0.066 Default 283011 0.009 0.094 0 1 83660 0.011 0.107 0 1 Loan amount 294664 901 2690 5 210000 84466 475 1160 5 93500 (CNY 10 thousand) Maturity Short term 294664 0.922 0.268 0 1 84466 0.983 0.129 0 1 Medium term 294664 0.06 0.237 0 1 84466 0.014 0.119 0 1 Long term 294664 0.019 0.136 0 1 84466 0.003 0.051 0 1 Loan type Secured loan 294664 0.446 0.497 0 1 84466 0.445 0.497 0 1 Fiduciary loan 294664 0.03 0.171 0 1 84466 0.018 0.134 0 1 loan on guarantee 294664 0.388 0.487 0 1 84466 0.426 0.495 0 1 Pledged loan 294664 0.079 0.27 0 1 84466 0.072 0.259 0 1 Discount loan 294664 0.056 0.23 0 1 84466 0.039 0.192 0 1 Firm size Micro and small firms 294664 0.833 0.373 0 1 84466 0.886 0.318 0 1 Medium firms 294664 0.13 0.337 0 1 84466 0.094 0.292 0 1 Big firms 294664 0.037 0.189 0 1 84466 0.02 0.14 0 1 Company age (Years) 294664 10.42 5.94 1 60 84466 11.23 4.87 1 37 Bank type Big five 294664 0.356 0.479 0 1 84466 0.297 0.457 0 1 Joint-stock 294664 0.137 0.344 0 1 84466 0.124 0.33 0 1 commercial banks City 294664 0.149 0.356 0 1 84466 0.1 0.3 0 1 commercial banks Rural banks 294664 0.358 0.479 0 1 84466 0.479 0.5 0 1 38 Table 11: Clean Air Action, default and loan spread, cluster by bank This table shows DID estimates of the e↵ect of the Clean Air Action on the default and loan spread of high polluting firms relative to low polluting firms. The dependent variable is default for columns (1) to (3) and loan spread for columns (4) to (6). Treat is a dummy variable marking all firms belonging to the high-polluting industries targeted by the Clean Air Action. Action is a dummy variable marking the post treatment period (6 January 2014 and 31 December 2014). All specifications contain loan, firm and macro-level controls. The lower part of the table denotes the type of fixed e↵ects. Standard errors are clustered at bank level and reported in parentheses (* p <0.10, ** p <0.05, *** p <0.01). Default Loan spread (1) (2) (3) (4) (5) (6) Action * Treat 0.0052*** 0.0053*** 0.0058*** 0.0127** 0.0123** 0.0115** (0.0017) (0.0017) (0.0017) (0.0049) (0.0049) (0.0043) Action 0.0026 -0.1256*** (0.0045) (0.0427) Treat -0.0007 -0.0007 -0.0012 0.0007 0.0009 0.001 (0.0016) (0.0016) (0.0015) (0.0044) (0.0043) (0.0038) Ln (Loan amount) -0.0000 -0.0001 -0.0000 -0.0100* -0.0100* -0.0100* (0.0008) (0.0008) (0.0008) (0.0051) (0.0051) (0.0051) Short-term loan 0.0099 0.0098 0.0046 0.047 0.0468 0.0403 (0.0125) (0.0125) (0.0103) (0.0405) (0.0405) (0.0379) Medium-term loan 0.0076 0.0076 0.0047 0.0071 0.007 0.0034 (0.0074) (0.0074) (0.0067) (0.0286) (0.0286) (0.0277) Fiduciary loan 0.0117 0.0117 0.0115 -0.0935*** -0.0936*** -0.0955*** (0.0082) (0.0082) (0.0080) (0.0250) (0.0251) (0.0256) Loan on guarantee -0.0037** -0.0038** -0.0038** -0.0806*** -0.0806*** -0.0809*** (0.0018) (0.0018) (0.0017) (0.0268) (0.0268) (0.0268) Pledged loan -0.0039 -0.0039 -0.0034 -0.1725*** -0.1724*** -0.1710*** (0.0034) (0.0034) (0.0034) (0.0438) (0.0438) (0.0445) Discount loan -0.0061 -0.0061 -0.0059 -0.0961*** -0.0961*** -0.0960*** (0.0042) (0.0043) (0.0042) (0.0318) (0.0318) (0.0324) Small and micro enterprises 0.006 0.006 0.0059 0.0707*** 0.0707*** 0.0709*** (0.0042) (0.0042) (0.0042) (0.0103) (0.0103) (0.0105) Medium enterprises 0.0018 0.0017 0.0018 0.0111** 0.0111** 0.0124** (0.0025) (0.0024) (0.0025) (0.0050) (0.0050) (0.0053) Firm age -0.0008*** -0.0008*** -0.0008*** -0.0004 -0.0004 -0.0004 (0.0002) (0.0002) (0.0002) (0.0011) (0.0011) (0.0011) Firm age Sq. 0.0000*** 0.0000*** 0.0000*** -0.0000 -0.0000 -0.0000 (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) (0.0000) Collective enterprises 0.0057** 0.0057** 0.0058** 0.0013 0.0014 0.0031 (0.0024) (0.0024) (0.0024) (0.0105) (0.0105) (0.0106) Private enterprises 0.0070*** 0.0070*** 0.0069*** 0.0139 0.0139 0.0153 (0.0021) (0.0021) (0.0020) (0.0116) (0.0116) (0.0115) Limited liability enterprises 0.0112*** 0.0112*** 0.0110*** 0.0018 0.0018 0.0032 (0.0022) (0.0022) (0.0021) (0.0112) (0.0112) (0.0112) Incorporated enterprises 0.0051** 0.0051** 0.0050** -0.0311** -0.0311** -0.0294** (0.0022) (0.0022) (0.0023) (0.0133) (0.0133) (0.0129) Joint venture enterprises 0.0110*** 0.0111*** 0.0108*** -0.0305** -0.0304** -0.0297** (0.0026) (0.0026) (0.0025) (0.0137) (0.0138) (0.0134) Foreign enterprises 0.0094*** 0.0094*** 0.0091*** -0.0539*** -0.0539*** -0.0523*** (0.0024) (0.0024) (0.0023) (0.0124) (0.0124) (0.0120) Other enterprises 0.0042* 0.0043* 0.0042* 0.0089 0.0089 0.0094 (0.0023) (0.0023) (0.0022) (0.0112) (0.0113) (0.0116) Share of secondary industry 0.8374** -2.0556 1.1906** 0.2416 -15.3263 8.696 (0.3241) (1.5747) (0.5496) (5.1644) (13.1714) (8.3422) Share of tertiary industry 0.8766** -2.6108 1.2632** 3.0117 -18.4133 9.8879 (0.3499) (2.1340) (0.5752) (6.3762) (16.7827) (9.8880) Ln (GDP per capita) -0.0800** 0.274 -0.1161** 0.7656* 1.8804 0.4334 (0.0377) (0.2262) (0.0541) (0.4263) (1.7584) (0.6816) Year fixed e↵ect Y Y Bank fixed e↵ect Y Y Y Y Prefecture fixed e↵ect Y Y Y Y Bank⇤year e↵ect Y Y Prefecture⇤year e↵ect Y Y Observations 366671 366671 366671 379130 379130 379130 R20.014 0.014 0.016 0.468 0.468 0.474 Adjusted R20.013 0.013 0.015 0.468 0.468 0.474 45 Table 12: Clean Air Action, default and loan spread, lagged by 30 days This table reports DID estimates on the e↵ect of the Clean Air Action on the default and loan spread where the implementation time of the Action is lagged by 60 days. The dependent variable is default for columns (1) to (3), and loan spread for columns (4) to (6). Treat is a dummy variable marking all firms belonging to the high-polluting industries targeted by the Clean Air Action. The lower part of the table denotes the type of fixed e↵ects. Standard errors are reported in parentheses (* p <0.10, ** p <0.05, *** p <0.01). Default Loan spread (1) (2) (3) (4) (5) (6) Action lag(30)*Treat 0.0048*** 0.0049*** 0.0054*** 0.0121*** 0.0116*** 0.0111*** (0.0008) (0.0008) (0.0008) (0.0014) (0.0015) (0.0015) Action lag(30) 0.0009 -0.1289*** (0.0024) (0.0049) Treat -0.0010* -0.0010* -0.0016*** 0.0008 0.0011 0.0009 (0.0005) (0.0005) (0.0006) (0.0010) (0.0010) (0.0010) Year fixed e↵ect Y Y Bank fixed e↵ect Y Y Y Y Prefecture fixed e↵ect Y Y Y Y Bank⇤year e↵ect Y Y Prefecture⇤year e↵ect Y Y Observations 342,665 342,665 342,665 360,059 360,059 360,059 R20.014 0.014 0.016 0.464 0.464 0.471 Adjusted R20.014 0.014 0.015 0.464 0.464 0.471 46 Table 13: Clean Air Action, default and loan spread, lagged by 60 days This table reports DID estimates on the e↵ect of the Clean Air Action on the default and loan spread where the implementation time of the Action is lagged by 60 days. The dependent variable is default for columns (1) to (3), and loan spread for columns (4) to (6). Treat is a dummy variable marking all firms belonging to the high-polluting industries targeted by the Clean Air Action. The lower part of the table denotes the type of fixed e↵ects. Standard errors are reported in parentheses (* p <0.10, ** p <0.05, *** p <0.01). Default Loan spread (1) (2) (3) (4) (5) (6) Action lag(60)*Treat 0.0053*** 0.0054*** 0.0062*** 0.0116*** 0.0110*** 0.0107*** (0.0008) (0.0008) (0.0008) (0.0015) (0.0015) (0.0015) Action lag(60) 0.0015 0.066 0.0061 -0.1319*** (0.0024) (0.0638) (0.0040) (0.0050) Treat -0.0009* -0.0010* -0.0015*** 0.0008 0.0011 0.0008 (0.0005) (0.0005) (0.0006) (0.0010) (0.0010) (0.0010) Year fixed e↵ect Y Y Bank fixed e↵ect Y Y Y Y Prefecture fixed e↵ect Y Y Y Y Bank⇤year e↵ect Y Y Prefecture⇤year e↵ect Y Y Observations 347,012 347,012 347,012 347,012 347,012 347,012 R20.013 0.013 0.015 0.462 0.463 0.470 Adjusted R20.013 0.013 0.014 0.462 0.463 0.470 47 Table 14: Clean Air Action, default and loan spread, multiple-period DID analysis This table reports the multiple period DID estimates on the e↵ect of the Clean Air Action on the default and loan spread. The dependent variable is default for columns (1) to (3), and loan spread for columns (4) to (6). Treat is a dummy variable marking all firms belonging to the high-polluting industries targeted by the Clean Air Action. Action is a dummy variable marking the post treatment period (6 Jan 2014 and 31 Dec 2014). Action1t, is a dummy that takes the value of 1 if a bank loan is granted during the interaction period (between 11 September 2013 and 5 January 2014), and 0 otherwise. The lower part of the table denotes the type of fixed e↵ects. Standard errors are reported in parentheses (* p <0.10, ** p <0.05, *** p <0.01). Default Loan spread (1) (2) (3) (4) (5) (6) Action1*Treat -0.0026** -0.0029*** -0.0031*** 0.0024 -0.0038** -0.0011 (0.0011) (0.0011) (0.0011) (0.0019) (0.0019) (0.0019) Action*Treat air 0.0057*** 0.0060*** 0.0063*** 0.0046*** 0.0113*** 0.0089*** (0.0008) (0.0008) (0.0008) (0.0015) (0.0015) (0.0015) Treat -0.0012** -0.0014** -0.0016*** 0.0115*** 0.0081*** 0.0082*** (0.0006) (0.0006) (0.0006) (0.0011) (0.0011) (0.0011) Action -0.0005 0.0031* 0.0031* 0.0159*** 0.005 0.0051 (0.0004) (0.0019) (0.0019) (0.0007) (0.0035) (0.0036) Action1 0.0025*** -0.0009 -0.0009 -0.0148*** -0.0042 -0.004 (0.0005) (0.0018) (0.0018) (0.0009) (0.0035) (0.0035) Year fixed e↵ect Y Y Bank fixed e↵ect Y Y Y Y Prefecture fixed e↵ect Y Y Y Y Bank⇤year e↵ect Y Y Prefecture⇤year e↵ect Y Y Observations 428,043 428,043 4280,43 450,13 450,138 450,138 R20.012 0.012 0.014 0.421 0.424 0.428 Adjusted R20.012 0.012 0.013 0.421 0.424 0.428 48 Table 15: Clean Air Action, default and loan spread, panel data analysis I This table reports DID estimates on the e↵ect of the Clean Air Action on the default and loan spread for a sample of firms that borrowed from the same bank both before and after the Action was implemented. The dependent variable is default for columns (1) to (3), and loan spread for columns (4) to (6). Treat is a dummy variable marking all firms belonging to the high-polluting industries targeted by the Clean Air Action. Action is a dummy variable marking the post treatment period (6 January 2014 and 31 December 2014). The lower part of the table denotes the type of fixed e↵ects. Standard errors are reported in parentheses (* p <0.10, ** p <0.05, *** p <0.01). Default Loan spread (1) (2) (3) (4) (5) (6) Action*Treat 0.0020*** 0.0021*** 0.0032*** 0.0081*** 0.0075*** 0.0069*** (0.0007) (0.0007) (0.0007) (0.0015) (0.0015) (0.0015) Action 0.0165*** -0.1015*** (0.0022) (0.0050) Treat 0.0018*** 0.0017*** 0.0009** 0.0044*** (0.0004) (0.0004) (0.0004) (0.0010) Year fixed e↵ect Y Y Bank fixed e↵ect Y Y Y Y Prefecture fixed e↵ect Y Y Y Y Bank⇤year e↵ect Y Y Prefecture⇤year e↵ect Y Y Observations 325437 325437 325437 325437 325437 325437 R20.016 0.016 0.019 0.464 0.464 0.471 Adjusted R20.016 0.016 0.019 0.464 0.464 0.471 49 Table 16: Clean Air Action, default and loan spread, panel data analysis II This table reports DID estimates on the e↵ect of the Clean Air Action on the default and loan spread for a sample of firms that borrowed from the same bank both before and after the Action was implemented. The dependent variable is default for columns (1) to (3), and loan spread for columns (4) to (6). Treat is a dummy variable marking all firms belonging to the high-polluting industries targeted by the Clean Air Action. Action is a dummy variable marking the post treatment period (6 January 2014 and 31 December 2014). The lower part of the table denotes the type of fixed e↵ects. Standard errors are reported in parentheses (* p <0.10, ** p <0.05, *** p <0.01). Default Loan spread (1) (2) (3) (4) (5) (6) Action*Treat 0.0012* 0.0013** 0.0023*** 0.0052*** 0.0050*** 0.0043*** (0.0006) (0.0006) (0.0006) (0.0015) (0.0015) (0.0015) Action 0.0150*** -0.0839*** (0.0023) (0.0052) Treat 0.0012*** 0.0012*** 0.0006* 0.0034*** 0.0035*** 0.0039*** (0.0003) (0.0003) (0.0003) (0.0011) (0.0011) (0.0011) Year fixed e↵ect Y Y Bank fixed e↵ect Y Y Y Y Prefecture fixed e↵ect Y Y Y Y Bank⇤year e↵ect Y Y Prefecture⇤year e↵ect Y Y Observations 273211 273211 273211 273211 273211 273211 R20.012 0.013 0.016 0.517 0.517 0.525 Adjusted R20.012 0.012 0.015 0.517 0.517 0.525 50 8 Appendix 8.1 Model Equations and First-Order Conditions 8.1.1 Households max E0 1 X t=0 ()tc1c t 1c vL ⌘(Lt)⌘, (8.1) subject to the following budget constraint: ct+dtwtLt+Rt1 ⇡t dt1+Ft.(8.2) The household optimality conditions yield the following firstorder conditions: {ct}:t=cc t(8.3) {dt}:t=EtRt ⇡t+1 t+1(8.4) {Lt}:vL(Lt)⌘1=wtt(8.5) 8.1.2 Entrepreneurs Re-write the entrepreneurs maximization problem by starting from the bank participation constraint: RL tbj e,t =((1µj)Z¯!j,t+1 0 !i j,t+1(1h)qj,k t+1⇡t+1kj e,tft+1(!i j)d!i j)+(Z1 ¯!j,t+1 Rj z,t+1bj e,tft+1(!i j)d!i j), (8.6) Define t+1(¯!j t+1)⌘¯!j t+1 Z1 ¯!j t+1 ft+1(!i j)!i j+Gt+1(¯!j t+1),(8.7) Gt+1(¯!bj,t+1)⌘Z¯!j t+1 0 !i bj,t+1ft+1(!i bj)d!i bj (8.8) and ¯!j t+1 =bj e,tRj z,t (qj,k t+1⇡t+1(1 k)kj e,t)(8.9) Use 8.7,8.8 and 8.9 to solve the bank participation constraint: RL tbj e,t =(1µj)(1h)qj,k t+1kj e,t⇡t+1(Z¯!j,t+1 0 !i j,t+1ft+1(!i j)d!i j)+Rj z,t+1bj e,t(Z1 ¯!j,t+1 ft+1(!i j)d!i j), 51 RL tbj e,t =(1µj)(1 h)qj,k t+1kj e,t⇡t+1(R¯!j,t+1 0!i j,t+1ft+1(!i j)d!i j) +¯!j t+1qj,k t+1⇡t+1(1 k)kj e,t(R1 ¯!j,t+1 ft+1(!i j)d!i j), RL tbj e,t =(1µj)(1 h)qk t+1kj e,t⇡t+1(R¯!j,t+1 0!i j,t+1ft+1(!i j)d!i j) +¯!j t+1qj,k t+1⇡t+1(1 k)kj e,t(R1 ¯!j,t+1 ft+1(!i j)d!i j+Gt+1(¯!bj,t+1)Gt+1(¯!bj,t+1)), RL tbj e,t =(1µj)(1h)qj,k t+1kj e,t⇡t+1Gt+1(¯!bj,t+1)+qj,k t+1⇡t+1(1k)kj e,t[t+1(¯!j t+1)Gt+1(¯!bj,t+1)], RL tbj e,t =(1h)qj,k t+1kj e,t⇡t+1[t+1(¯!j t+1)µjGt+1(¯!bj,t+1)] (8.10) Re-call the budget constraint: cj e,t +Xt+qj,k t(kj e,t (1 k)kj e,t1)+wj tLj t+RK tkj e,t +[1Fj,t(¯!j t)]Rj z,t1bj e,t1 =Yj e,t +bj e,t qj,k t(1 k)kj e,t1Gj t(¯!j t), and substitute the threshold value: cj e,t +Xt+qj,k t(kj e,t (1 k)kj e,t1)+wj tLj t+RK tkj e,t +[1Fj,t(¯!j t)]¯!j t+1qj,k t+1⇡t+1(1 k)kj e,t =Yj e,t +bj e,t qj,k t(1 k)kj e,t1Gj t(¯!j t), Let’s define [1 Fj,t(¯!j t)]¯!j t+1 =[ t+1(¯!j t+1)Gt+1(¯!bj,t+1)],and substitute in the previous expression: cj e,t +Xt+qj,k t(kj e,t (1 k)kj e,t1)+wj tLj t+RK tkj e,t +[ t+1(¯!j t+1)Gt+1(¯!bj,t+1)]qj,k t+1⇡t+1(1 k)kj e,t =Yj e,t +bj e,t qj,k t(1 k)kj e,t1Gj t(¯!j t), cj e,t +Xt+qj,k t(kj e,t (1 k)kj e,t1)+wj tLj t+RK tkj e,t +[ t+1(¯!j t+1)Gt+1(¯!bj,t+1)+µjGt+1(¯!bj,t+1)µjGt+1(¯!bj,t+1)]qj,k t+1⇡t+1(1 k)kj e,t =Yj e,t +bj e,t qj,k t(1 k)kj e,t1Gj t(¯!j t), Finally, we obtain: cj e,t+Xt+qj,k t(kj e,t(1k)kj e,t1)+wj tLj t+RK tkj e,t+RL tbj e,t =Yj e,t+bj e,tµjGj t(¯!j t)qj,k t(1k)kj e,t1, 52 Re-write the maximization problem as: max E0 1 X t=0 (e)thln(cj e,t)i(8.11) subject to: cj e,t+Xt+qj,k t(kj e,t(1k)kj e,t1)+wj tLj t+Rj,K tkj e,t+RL tbj e,t =Yj e,t+bj e,tµjGj t(¯!j t)qj,k t(1k)kj e,t1, bj e,t mj e,t Et (qj,k t+1⇡t+1(1 k)kj e,t) RL t ,(8.12) and Mt⌦ng,t "M,t (8.13) where Yj e,t =At(kj e,t1)↵(Lj t)1↵jXj t.(8.14) The entrepreneurs’ optimality conditions yield the following firstorder conditions: {cj e,t}:j e,t =(cj e,t)c(8.15) {Lj t}:wj t=(1↵)Yj e,t Lj t (8.16) {ke,t}:qj,k tj e,t =eEt(Rj,K t+1 +(1k)qj,k t+1(1 µGj t(¯!j t))j e,t+1))+ Et(⇤j e,t+1mj e,t+1 (qj,k t+1⇡t+1(1k)) (RL t1))(8.17) {¯!j t+1}:e⇤j e,t+1µj@Gj t+1(¯!j t+1) @¯!j t+1 =j e,t+1⇡t+1 @mj e,t+1 @¯!j t+1 (8.18) {Eg t}:1=(g)Yj e,t Et (8.19) {Mng t}:1=(ng)Yj e,t Mt + t(8.20) where j e,t is the lagrangian multiplier on entrepreneurs budget constraint, ⇤j e,t is the lagrangian multiplier on the participation constraint and tis the lagrangian multiplier on the pollution constraint. 8.1.3 Banks {dt:}1 cb,t [1 d]=bEt(Rt cb,t+1⇡t+1 )+µb t(8.21) 53 {bt:}1 cb,t [1 + b]=bEt(RL t cb,t+1⇡t+1 )+(1)µb t(8.22) 54