Modeling the impact of emission credit systems on automotive product portfolios: A mathematical analysis of policy effects in Europe, China, and the U.S. under different demand scenarios

Shi, Zewei Article Modeling the impact of emission credit systems on automotive product portfolios: A mathematical analysis of policy effects in Europe, China, and the U.S. under different demand scenarios Junior Management Science (JUMS) Provided in Cooperation with: Junior Management Science e. V. Suggested Citation: Shi, Zewei (2025) : Modeling the impact of emission credit systems on automotive product portfolios: A mathematical analysis of policy effects in Europe, China, and the U.S. under different demand scenarios, Junior Management Science (JUMS), ISSN 2942-1861, Junior Management Science e. V., Planegg, Vol. 10, Iss. 3, pp. 748-780, https://doi.org/10.5282/jums/v10i3pp748-780 This Version is available at: https://hdl.handle.net/10419/326972 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/4.0/ Junior Management Science 10(3) (2025) 748-780 Junior Management Science www.jums.academy ISSN: 2942-1861 Editor: DOMINIK VAN AAKEN Advisory Editorial Board: FREDERIK AHLEMANN JAN-PHILIPP AHRENS THOMAS BAHLINGER MARKUS BECKMANN SULEIKA BORT ROLF BRÜHL KATRIN BURMEISTER-LAMP CATHERINE CLEOPHAS NILS CRASSELT BENEDIKT DOWNAR KERSTIN FEHRE MATTHIAS FINK DAVID FLORYSIAK GUNTHER FRIEDL MARTIN FRIESL FRANZ FUERST WOLFGANG GÜTTEL NINA KATRIN HANSEN ANNE KATARINA HEIDER CHRISTIAN HOFMANN SVEN HÖRNER STEPHAN KAISER NADINE KAMMERLANDER ALFRED KIESER ALEKSANDRA KLEIN NATALIA KLIEWER DODO ZU KNYPHAUSEN-AUFSESS SABINE T. KÖSZEGI ARJAN KOZICA CHRISTIAN KOZIOL MARTIN KREEB WERNER KUNZ HANS-ULRICH KÜPPER MICHAEL MEYER JÜRGEN MÜHLBACHER GORDON MÜLLER-SEITZ J. PETER MURMANN ANDREAS OSTERMAIER BURKHARD PEDELL ARTHUR POSCH MARCEL PROKOPCZUK TANJA RABL SASCHA RAITHEL NICOLE RATZINGER-SAKEL ASTRID REICHEL KATJA ROST THOMAS RUSSACK FLORIAN SAHLING MARKO SARSTEDT ANDREAS G. SCHERER STEFAN SCHMID UTE SCHMIEL CHRISTIAN SCHMITZ MARTIN SCHNEIDER MARKUS SCHOLZ LARS SCHWEIZER DAVID SEIDL THORSTEN SELLHORN STEFAN SEURING VIOLETTA SPLITTER ANDREAS SUCHANEK TILL TALAULICAR ANN TANK ANJA TUSCHKE MATTHIAS UHL CHRISTINE VALLASTER PATRICK VELTE CHRISTIAN VÖGTLIN BARBARA E. WEISSENBERGER ISABELL M. WELPE HANNES WINNER THOMAS WRONA THOMAS ZWICK Volume 10, Issue 3, September 2025 JUNIOR MANAGEMENT SCIENCE Johannes Witter,Predicting Stock Returns With Machine Learning: Global Versus Sector Models Robin Roskosch, Beware of Bullshit –A Qualitative Study on Young Adults’ Sustainability Awareness of Online Services Nadhilla Mazaya,Board Gender Diversity: Evidence From Indonesia Alexander Sake, Value Creation Opportunities of Generative AI –A Case Study Justus Olbrich, The Effect of Changes in Internal Control Systems on Audit Risk Jan Oliver Horstmann, Mandatory ESG Disclosure and Firm Value –A Quantitative Analysis of the Effect of Directive 2014/95/EU on Firm Value Meret Anna Gläser, Government Interventions During the COVID-19 Pandemic, Culture, and Corporate Cost Behaviour Zewei Shi, Modeling the Impact of Emission Credit Systems on Automotive Product Portfolios: A Mathematical Analysis of Policy Effects in Europe, China, and the U.S. Under Different Demand Scenarios Hagen Alexander Hönerloh, Numerical Studies for the Scheduling of Continuous Annealing Lines Lea Wedel, KPIs for Sustainability: Defining the Strategy for a Sustainable Future in the Insurance Industry 561 582 609 631 657 677 715 748 781 810 Published by Junior Management Science e.V. This is an Open Access article distributed under the terms of the CC-BY-4.0 (Attribution 4.0 International). Open Access funding provided by ZBW. ISSN: 2942-1861 Modeling the Impact of Emission Credit Systems on Automotive Product Portfolios: A Mathematical Analysis of Policy Effects in Europe, China, and the U.S. Under Different Demand Scenarios Zewei Shi Technical University of Munich Abstract In the midst of the global climate crisis, governments worldwide have implemented a range of emission policies aimed at encouraging more production of the environmentally friendly vehicle. However, the exact impact of these policies on automakers’ production portfolios and profitability remains uncertain and challenging to anticipate. This paper presents a comprehensive analysis of three major emission regulation policies enacted by the European Union (EU), China, and the United States (U.S.), evaluating their influence on car manufacturers. Leveraging a mathematical model, this paper adopt the perspective of individual manufacturers seeking to maximize revenue, delving into the intricacies of these policies. Furthermore, this article conduct sensitivity and factorial analyses to assess the impact of policy parameters. The findings reveal that all three major emission policies contribute to an increase in the production of low-emission vehicles. However, China’s policy has the least impact on manufacturers’ profits and relies more on market demand to reduce the average carbon fleet emissions compared to the policies in the EU and the U.S. In conclusion, this paper underscores that different policy systems yield varying profit outcomes for manufacturers, necessitating adjustments to production portfolios for sustained profitability and the significance of mathematical models in aiding manufacturers’ understanding of evolving policies and making informed predictions in a dynamic regulatory landscape. Keywords: automotive production; green transition; international emission policies; regulatory impact; sustainability 1. Introduction As modern industrialization surges forward, humanity confronts the complex challenges of climate change. This encompasses the onset of extreme weather patterns and elevated temperatures, both driven by the incessant release of copious amounts of greenhouse gases into the atmosphere. The excessive emissions of greenhouse gases, such as carbon dioxide (CO2)and methane (CH4), instigate the greenhouse effect, culminating in the far-reaching issue of global I would like to sincerely thank my supervisor, Maximilian Kolter for his thoughtful guidance and support throughout my thesis. He not only helped me approach the problem more effectively but also provided valuable feedback to improve the clarity and readability of my writing. I am also grateful to Professor Kolish for providing the data and initial literature that laid the groundwork for my research. warming. This phenomenon poses a threat to the existing ecosystem, manifesting in disruptive weather patterns and extreme climatic events (Yoro & Al., 2020). Notably, the primary source of CO2emissions stems not only from industrial production but also from vehicular exhaust (Huang et al., 2015). Traditional vehicles predominantly powered by gasoline and diesel generate substantial CO2emissions in dayto-day usage. In response to this environmental challenge, the electric vehicle concept emerges as a viable solution. By utilizing electricity as the primary power source, electric vehicles could reduce carbon emissions, positioning them as a more eco-friendly alternative (Costa et al., 2021) Currently, there are four main types of vehicles: Internal Combustion Engine Vehicles (ICEV), Plug-in Hybrid Vehicles (PHEV), Battery Electric Vehicles (BEV), and Fuel Cell ElecDOI: https://doi.org/10.5282/jums/v10i3pp748-780 © The Author(s) 2025. Published by Junior Management Science. This is an Open Access article distributed under the terms of the CC-BY-4.0 (Attribution 4.0 International). Open Access funding provided by ZBW. Z. Shi /Junior Management Science 10(3) (2025) 748-780 749 tric Vehicles (ICEV) (Vacheva & Hinov, 2019). ICEV cars are the most traditional, powered by gasoline and emitting a relatively high amount of CO2. BEV and FCEV are pure electric vehicles with zero carbon emissions, collectively referred to as ZEV (Zero Emission Vehicles). The distinction between these two vehicle types lies in their power sources: BEVs use batteries charged with electricity, while FCEVs utilize biofuels and hydrogen-powered fuel cells, which can also be considered hydrogen-powered vehicles (Parikh et al., 2023). PHEVs represent a middle ground, with lower carbon emissions compared to ICEVs. They operate on a hybrid power source, utilizing both electricity and petrol or gasoline simultaneously. These different vehicle types have significant variations in production costs and sales revenue, depending on factors such as production year, type, and size (Lipman & Delucchi, 2006). The ZEV and PHEV vehicle types produce less CO2 with lower tailpipe emissions, which can help mitigate the problem of global warming. However, despite the increasing popularity of electric and low-carbon emission vehicles, the cost of new energy vehicles remains higher than traditional vehicles, leading to lower profits for manufacturers (Cuenca et al., 2000). Vehicle producers need to maintain market competitiveness and prioritize profit and financial gains, making it challenging to persuade them to prioritize the production of less profitable but environmentally friendly vehicles. To encourage manufacturers to reduce fleet emissions, which represent the average amount of CO2 produced across their production portfolio, various governments have introduced policies and regulations. Different governments have adopted unique approaches to establish country-specific policies (An et al., 2011). However, the precise effectiveness of these policy systems on manufacturers’ production portfolios remains unclear. This article focuses on three primary markets: Europe, China, and the USA, each with its own emission regulations. In Europe, the emission policy is referred to as the Super Credit Policy, in China, it is known as the Dual Credit Policy, and in the USA, it is named the US Credit Policy. Further details about these regulations can be found in Section 3.4. In this scenario, two stakeholders exist: the manufacturer and the policymaker. For the manufacturer, they need to determine quickly how different policies will affect their business profit and make quick responses to the production portfolio to ensure profitability. For the policymaker, they need to balance the profit of the manufacturer and the carbon emission. As Dominioni and Faure (2022) shows, an imbalanced policy can lead to either a loss of tax income or the limited effectiveness of emission policies. Although the government has spent a lot of time discussing the details of the policy and understands that even small parameter changes can affect policy effectiveness, evaluating the policy’s effect is complex. Policies must precede market reactions, and past results are unreliable predictors due to changing demand and production situations. Once a policy is published, it is difficult to withdraw. For manufacturers, regardless of how the policy performs, they need to understand how different policies and changes in policy parameters will affect their production portfolio and revenue to adjust their business strategy. Different countries have their own goals in setting up policies, and it is essential for manufacturers to analyze how different policies in different countries differ to differentiate their production strategy and ensure better profitability. Overall, this study aims to compare various emission policies and assess their impact on manufacturing portfolios for producers across different parameter scenarios in a quantitative manner. In this article, operations research methods are used to simulate the effect of different policies based on realistic test instances. By constructing mixed-integer linear optimization models for different policies and demand scenarios, policymakers as well as vehicle manufacturers can have a detailed quantitative view to assess policy effectiveness (Thies et al., 2022). The data comes from large global vehicle manufacturer, and the policy information is based on the current policy setup as of 2023. Moreover, since the mathematical model is flexible in changing parameters, the parameters and data can be adjusted to reflect the current situation and make more accurate predictions and analyses. The remainder of this study is structured as follows: In Section 2, I discuss the literature related to this topic on finding out the emission policy impact. In Section 3, I present the model with detailed formulation and parameters, along with different models of the three emission policy systems in Europe, China, and the US. In Section 4, I show how the model is solved for different policy systems, with the main discussion focusing on the Super Credit System in Europe. In Section 5, I list out the test instances performed and the structure of the design of the experiment in evaluating different policy systems. In Sections 6, I present the final results, including the detailed portfolio, as well as sensitivity and factor analysis for different demand scenarios and parameter settings. Finally, in Section 7, I draw conclusions regarding the different policy systems and provide an outlook for further research. 2. Literature Review Various approaches have been proposed to assess the effects of emission policies on car manufacturers, and they can be classified into five primary categories. Empirical studies and economic models rely on historical data and economic principles to conduct analyses on a broad scale. Market scenario models create hypothetical market conditions to evaluate policy impacts. A technology strategy model employs mathematical modeling to assist manufacturers in making decisions regarding the adoption of different vehicle models with various motive technologies. Simulation-based planning models use simulations to project long-term effects of the policy regulation for the manufacturer. Individual vehicle manufacturer mathematical models are tailored to specific manufacturers for in-depth analysis with output of detailed production plan and fleet emission trend. In the category empirical studies and economic models, with the empirical studies gather data on emissions, produc- Z. Shi /Junior Management Science 10(3) (2025) 748-780750 tion, and market behavior to analyze the impact of emission policies and the economic model use economic principles to predict how policy changes may affect car manufacturers, such as changes in costs, prices, and market demand. For the empirical studies Bergek and Berggren (2014) reviewed the empirical studies on environmental policy and found that policy instruments play a crucial role in driving environmental innovation across sectors. Also, (Y. Wang et al., 2018) analyzed compliance strategies of four different automakers under dual-credit regulations, considering fuel economy and NEV (New Energy Vehicle) production which includes the ZEVs (Zero Emission Vehicles) and PHEVs (Plug-in Hybrid Electric Vehicle), comparing their approaches and suggesting cost-effective strategies with regulatory improvements. Besides these empirical analyses, the economic model and pricing model have been built up to analyze how the policy would affect the vehicle manufacturer in a broad vision. (Moran et al., 2020) conducted micro-level studies with a multi regional input-output economic model to analyze the consumer-oriented policy and showed that these policies would reduce carbon emissions by about 25%. Additionally, the government pricing model of dual-credit policy published by (Yang et al., 2023), which compared with the market pricing model, shows that the dual-credit policy benefits energy saving and emission reduction in the transport sector. Moreover, in the study by (Ma et al., 2021), a supply chain model includes two stakeholders, the engine supplier and automakers, to analyze the carbon emission policy effect on the production of the ICEV and NEV vehicle. (Michalek et al., 2005) also considered the impact of the competition of other manufacturer in the paper and proposed mathematical models of engineering performance, consumer demand, and manufacturing costs, combined with game theory for the market segment. For these models, the trend could be observed from the economic perspective, but it could be too broad in scope that makes lack of some precision in explaining some details in the impact of the emission policy on the production plans of the individual manufacturer. Besides the economic models, various market scenario models have been used to analysis the impact of the emission policy on the manufacturer’s portfolio. These models creating various hypothetical market scenarios including the transportation sector and then evaluating how different emission policies would impact car manufacturers under these scenarios. For example, Thiel et al. (2016) used a TIMES-based energy system model to examines the impact of stricter EU CO2 car legislation on transport-related emissions, Electric vehicle uptake, oil consumption, and energy costs. This model is a modeling platform that consider factors like energy production, costs, and environmental impact and could helps make informed decisions about emission policies and resources. Hill et al. (2018) provided three models of PRIMES (global energy-economic model)-TREMOVE (transportation policy), GEM-E3T (model with macroeconomic, energy, and environmental policies) and the JRC DIONE (model for assessing energy and environmental policies) to analysis the overall market situation in considering energy, climate, transportation and the Europe emission policy. These models all reach similar conclusion that the EU policy are effective in reducing GHG(Greenhouse Gas) emissions. The ALTER-MOTIVE modelling method also been conducted by Ajanovic and Haas (2017) to integrate the energy system and transportation showing that GHG emissions could be reduced at least by 33% in a selected policy scenario. Other than these pre-formed model, Ou et al. (2018) develops the New Energy and Oil Consumption Credits Model to quantify the impacts of this policy in scenario from 2016 to 2020 to discuss the effect of the dual credit system in China on the electric vehicle sales. While these scenario models are useful for generating convincing results, they may oversimplify market conditions and the behaviors of individual manufacturers. These models can assess the effects of emission policies on a broad scale and from a market perspective, but they may not provide a comprehensive understanding of how individual car manufacturers would be impacted by or respond to these policies. Speak to make analysis from individual level, there are some studies modeled the problem of individual car manufacturers to find profit maximizing technology strategies considering emission policies. S. Wang et al. (2018) build a mixed-integer mathematical model with decision variables representing various motor technologies in a technology combination (TC) problem. This model is designed to describe an automaker’s decision-making process, and I utilize a genetic algorithm to assess the impact of China’s dual credit policy from 2020 to 2025. Moreover, Romejko and Nakano (2017) increase the range of the motor technology path for the vehicle projects and explores a more diverse range of eight alternative fuel vehicles (AFVs), including EVs, FCVs, CNG (Compressed Natural Gas) vehicles, and more, to predict an optimal AFV portfolio for achieving economic and energy security goals. Moreover Zhu et al. (2022) proposed a decision-making algorithm for automakers’ production strategies under the dual-credit policy in China. This algorithm reveals how automakers’ choices transition between strategies based on thresholds and government targets. Furthermore, other than changing between the constraints and parameter, Kellner et al. (2021) also use multi-objective optimization method analysis in the technology selection to find the optimal power train technology portfolio. For these models, some individual level of the vehicle portfolio planning has been performed. However, these models only provide the outcome of which motor technology to initialize, without offering information on the actual quantity produced for different vehicle technology types. Consequently, they do not offer a comprehensive view of the resulting average fleet emission values or the detailed cost structure of vehicle manufacturers and are not capable for providing a very clear picture of the vehicle portfolio and the impact of emission policy on these portfolios. Simulation method is also a popular modeling method to obtain vehicle portfolios. Kieckhäfer et al. (2012) at year 1970 used a hybrid market simulation approach for strategic planning of automotive vehicle portfolios to predict power- Z. Shi /Junior Management Science 10(3) (2025) 748-780 751 train market shares across vehicle sizes based on portfolio offerings, consumer behavior, and market conditions. Also Kieckhäfer et al. (2009) creates a framework using system dynamics and agent-based simulation to analyze the product strategies which aid manufacturers in effective technology introductions across vehicle classes while considering regulations and markets. Moreover the NW (Newman and Watts) ’small world’ network model also been used by Hu et al. (2020) to explore the dynamic effects of different policies on the diffusion of electric vehicles. For these simulation model, they could provide a close to reality output for the vehicle portfolio result thus indicate the impacts of the emission policy, but the model would take long times to be solved. It could be inefficient in time when implementing large number of testing scenarios into those simulation model for getting the solution insights. Recently, Thies et al. (2022) proposed a novel model that concentrates on the individual vehicle manufacturing portfolio. This article analysis how the EU emission policy effects the vehicle portfolio using optimization model from the perspective of an individual vehicle manufacturer. This model has provided a detailed production portfolio, showing the optimal quantity of each vehicle type to be produced annually, along with a comprehensive resource plan and accurate average fleet emissions for each year’s production and provide a basic framework of the optimization model used in this article. In this article, the base model of Thies et al. (2022) is extended by considering not only the EU policy but also the other two major policy system in China and the US. Furthermore, the EU policy is considered in more detail, as the super credit relaxation is considered. This novel model offers a more comprehensive and realistic analysis of policy effects from the perspective of individual car manufacturers and can be solved in few minutes. This model provides detailed information, including the vehicle initialization plan, production quantity for each vehicle type in each year, average fleet emissions in each year, market share of vehicle types, and the detailed cost structure. Moreover, it allows for easy parameter adjustments to customize the results as the policy changes. 3. Method This section presents a mathematical model for project portfolio planning considering different emission policies. The overall road map for the model formulation is presented in Section 3.1 and Section 3.2 explains all the parameter information for the mathematical model. After that, a base model without emission policies is presented in Section 3.3, before it is extended by the emission policies for Europe, China and the U.S. in Section 3.4. 3.1. Model Road Map Figure 1describes the roadmap for the optimization model in this article. The planning horizon is 10 years, from 2025 to 2035, during which the manufacturer can change its portfolio production decisions. For the period outside the planning period, the settings remain fixed. Within the planning horizon, the manufacturer can make several decisions, with the primary one being the determination of production quantities (qvt )for each vehicle (v)in each period (t). To ensure these production quantities, the resource plan has also been established, which includes the capacity (kr t )for each resource (r)in each period (t). Additionally, the resource adjustment plan contains (kRampup rt ) and (kRampdown rt ), representing the required increase and decrease amounts for resource (r)in period (t). The objective function aims to maximize the Net Present Value (NPV) for the entire production portfolio, which is discounted the manufacturer’s profit obtained by subtracting all costs from the revenues. Revenues encompass sales revenue generated from vehicle sales throughout the planning period, as well as end-of-period capacity cash-out income. Costs include production expenses, which consist of both fixed and variable costs related to vehicle production, expenses for increasing or decreasing resources to maintain production capacity, development costs for initializing new vehicle projects, and penalty costs or gains dependent on the specific policy system in place. Several constraints bind the decision-making process for vehicle production and can be categorized into five main categories. The vehicle project constraints help determine the initialization of different vehicle projects each year and limit the maximum number of projects that can be started. The production resource constraints are used to ensure sufficient resources are available for production. The vehicle demand constraint is employed to prevent the sale of more vehicles than the market demand, and the production volume constraint is used to ensure the minimum production volume each year. The policy-related constraints in different policy systems impose penalties or restrictions on the average fleet emissions of the production portfolio. The inputs for modeling the production process constraints are depicted in Figure 1. The vehicle projects include several pre-defined projects categorized by powertrain technology type, size, production year, and power range class. Some vehicle projects are already determined before the planning horizon. Each vehicle project has its own production cost, sales revenue, and tailpipe line emission value. Development costs are assumed to be the same for all vehicle projects, and the maximum life cycle is equal for all projects. After 2025, within the planning period, the manufacturer can initialize new vehicle projects if suitable.Each vehicle project requires production resources. Before the planning period, the current on-hand resources are pre-defined. The manufacturer must ensure that resources can meet the production quantity of vehicle projects, which includes decisions on ramping up or down capacity. Resource costs include fixed and variable costs, with the fixed cost per production resource potentially decreasing due to economies of scale. Sales quantities must not exceed the market demand for a given year, segmented by vehicle type, size, and power range Z. Shi /Junior Management Science 10(3) (2025) 748-780752 Figure 1: Optimization Model Road Map (e.g., IC EV_medium_low). Multiple vehicle projects can be considered for several years within a market demand segment, but each vehicle project is associated with only one market segment. Quantities below market demand are assumed to be sold, and any unfilled demand is considered to be lost with no capacity payback. The general settings define the framework for general parameters, such as the assumed annual interest rate for NPV calculation, utilization loss due to capacity increment, and the minimum production quantity required to stay on the market, with full details described in section 3.2.3. In addition to the basic setup, one of the policies from Super Credit Policy in Europe, Dual Credit Policy in China, US Emission Policy in the United States, or no Policy should be chosen to form the final policy-specific optimization model. Each policy entails specific parameters and formulas to be considered. The parameter values for different policies are based on current research in 2023, combined with personal assumptions, and all are related to the CO2average fleet emissions, as described in section 3.4. Once these setups are incorporated into the model and the model is solved, the final objective value becomes available, along with detailed values for revenues and various cost sectors. Additionally, the fleet emissions for each vehicle project (v)in each period (t)can be determined by multiplying the carbon cycle emissions (Ev)by the quantity produced (qvt )each year. All details about the formulation of the optimization model are presented in section 3.3. 3.2. Parameter Information for the Vehicle Project Portfolio Planning This section provides detailed parameter information, including sets, indices, decision variables, and descriptions for all parameters used in the model. 3.2.1. Sets & Indices The model uses several sets and indices, which are described in Table 1. These sets and indices are essential for defining the parameters and decision variables in the model. 3.2.2. Decision Variables There are 11 sets of decision variables described in Table 2. Three of these variables are binary variables, while Z. Shi /Junior Management Science 10(3) (2025) 748-780 753 Table 1: Sets and Indices Set & Indices Set Index Description VνVehicle projects X∈VVehicle project start before the planning horizon T t Periods P∈TPeriods in planning horizon M m Market segments R r Production Resources E∈RExisting production resources at the beginning of the planning horizon the rest are continuous variables. These decision variables play a critical role in the model, enabling the optimization of the vehicle portfolio and the assessment of various policy impacts. 3.2.3. Parameters Table 3concludes all the parameters used for the mathematical model, including the basic model as well as the additional emission policy model and separated by categories. The parameter contains type constant, variable, and vector. 3.3. Mathematical Formulation for the Base Model This section describes the model for the base model without the emission policy but provides an explanation of the objective function and the constraints in the context of a pure vehicle manufacturer setting. 3.3.1. Objective Function For the objective function, the goal is to maximize the net present value (NPV) of the vehicle project in ten years period. The interest rate is assume to be 5%, and the monetary value is calculated in each year and return the final NPV value base on year 2025. The objective function consists of three major parts which are the net profit of the production portfolio, the development cost and capacity increment cost, and finally the capacity cash back cost calculated and the end of the production planning phase at year 2035. For the net profit of each year in the planning period, it adds up all sale revenue according to the production quantity and minus the production variable cost as well as the fixed cost, also the penalty cost due to the emission policy would be deducted according to different policy types. For the development cost and capacity increment cost although the cost are spent on specific year in the production planning period, but it is assumed that the payment does not due immediately. The payment and deduction could be evenly distributed in 7 years period with each year about 14.3% of the total cost paid. VectRD vt is the parameter used for calculate the distribution of the cost and could also be counted after the production period. So the net NPV is summation for a all periods which is t∈T, from 2025 to 2050. For the third part of the capacity back value, it is calculated at the end of the planning period at year 2035. This term used to count back the capacity value to prevent the model over estimated the cost for the last few years capacity increase. For the capacity back cost, it is assumed that the resource on hand would retain it’s value in 10 years period and for each year, the value would depreciate by 10%. For example, for the resource in 2030, the resource would be capacity back with the rest value of 50%. Objective Function: Maximize NPV with: maxNPV = X t∈P ((T CSaleProd t−T C ProdFixed t−T C Penalt y t)·drt) −X t∈T ((T CRD t+T CCapacit y t)·drt)+TRCapaBack·drTmax (1) with T CSaleProd t=X v∈V ((sunit v−cUnitvar v)·qvt )(2) T C ProdFixed t=X r∈R (cResource ·kr t )(3) T C Penalt y t=cPenalt y t(4) T CRD t=X v∈V (cRD v·VectRD vt ·yv)(5) T CCapacit y t=X r∈R X t∈P (cRampU pF ixed ·VectPC τt·yRampupBin rt +cRampU pVar ·VectPC τt·kRampup rt )(6) TRCapaBack =X r∈R zRestValue r(7) 3.3.2. Constraints From the baseline model, there are four categories of constraint sets listed below. The constraints for different emission policies are in Section 3.4. For the vehicle projects constraints set, it consists of constraints related to vehicle project initialization. Constraint (8) sets up the initial start for the vehicle project before 2025, the planning horizon. Constraint (9) uses a big number M to switch the binary variable for the start of the new vehicle project. Constraint (10) limits the allowed start for the number of vehicle projects in every year due to resource limits, and Constraint (11) forces the quantity produced to 0 if the vehicle exceeds the project life cycle of tmax years. Z. Shi /Junior Management Science 10(3) (2025) 748-780754 Table 2: Description of Decision Variables Decision Variables Variable Type Range Description yvBinary {0,1}Vehicle project starting indicator with 1 meaning vehicle vis realized and 0 otherwise qvt Continuous R+Number of vehicle project vproduced at time period τ krt Continuous R+Capacity of resource rin period t kRampup rt Continuous R+Increase of production resources rat the beginning of period t kRampdown rt Continuous R+Decrease of production resources rat the beginning of period t yRampup rt Binary {0,1}Resource increase indicator with 1 meaning there is an increase in resource rat the beginning of period t, with 0 otherwise zRestValue rContinuous R+Residence Value for resource rat the end of the planning horizon cPenalt y tContinuous R+Penalty cost paid for the excess CO2Emission in period t Prelax tContinuous R+Super credit policy relaxed percentage y1Binary {0,1}Binary variable used to form maximum or minimum constraint DDual tContinuous R+Cost paid(+) or Revenue earned (-) for dual credit policy For the production resource constraints, Constraint (12) describes the resource usage constraint, and Constraint (13) indicates that in the ramp-up period, the resource would only be available at θmax percentage. Constraint (14) adjusts the on-hand resource at each year of the planning period after the previous ramp-up and ramp-down decisions that would be made. Constraint (15) switches on the fixed cost for ramping up the capacity. Constraint (16) calculates the rest value of the on-hand resources at the end of the planning year 2035, with Vres representing the remaining value, estimated as 5% of the total cost paid for increasing this amount of resource. Constraint (17) is the vehicle demand constraint to ensure that the production volume, which is less than the demand, would be sold to earn profit. Constraint (18) guarantees the minimum production volume in each year in the planning period. Vehicle Projects yv=yinitial v∀v∈X(8) qvt ≤M·yv∀v∈X,∀t∈P(9) X v∈V:SOPv=t yv≤SOPmax ∀t∈P(10) qvt =0∀v∈V,∀t∈T:t≤SOPv ∨t≥SOPv+tmax (11) Production resources X v∈V:rv=R qvt ≤kr t ∀r∈R,∀t∈P(12) X v∈V:rv=R qvt ≤θmax ·kr t ∀r∈R,∀t∈P:t=tRampup r(13) krt =kinitial r+X τ∈P:τ≤t (kRampup rτ−kRampdown rτ) ∀r∈R,∀t∈P(14) kRampup rt ≤M·yRampup rt ∀r∈R,∀t∈P(15) zRestValue r≤Vres ·X t∈U ((yRampup rt ·cRampU pF ix ed ) + (cRampU pVar ·kRampUp rt )) ∀r∈R(16) Vehicle Demand X v∈V:mv=m qvt ≤dmt ∀m∈M,∀t∈P(17) Production volume X v∈V qvt ≥qmin ∀t∈P(18) Z. Shi /Junior Management Science 10(3) (2025) 748-780 755 Table 3: Detailed Parameter Information Parameter Parameter Category Parameter Type Description Vehicle Projects SOPvVector Start of production time for vehicle project v cUnitvar vConstant Unit variable cost for vehicle project v sUnit vConstant Unit sale revenue for vehicle project v EvVector CO2cycle emission of vehicle project v mvVector Market segment type of vehicle project v rvVector Resource type needed for vehicle project v tmax Constant Maximum duration of the selling time period VectRD vt Vector Distribution of cash flow: Percentage of total development cost for vehicle project vin period t cRD vConstant Development cost for a new vehicle model yinitial vVector Preset vehicle realization indicator for vehicle project v Production Resource cRampUpFixed Constant Fixed cost for ramping up production resources cRampUpVar Constant Unit variable cost for the production resources tRampup rVector Ramp-up period for production resource r VectPC tcur trampup Vector Distribution of cash flow: Percentage of total production cost for ramp-up period tcur distributed in current period trampup cResource Constant Constant cost for each unit of production resource on hand kinitial rVector Capacity of production resource rbefore the planning horizon Vres Constant Residence value for production resource at the end of planning horizon Demand dmt Vector Number of vehicles demanded in market segment mat period t Pmin Constant Minimum percentage of total demand needed to be fulfilled Emission Regulation ELaw tConstant Total CO2fleet emission threshold in period t σlow Constant The maximum fleet emission value for the category of low emission vehicles in the super credit policy system γmutiplier Constant PHEV multiplier factor in super credit policy system ERelax Law tVariable Relaxed threshold for total CO2fleet emission in period t PEV thres tVariable The regulated electric vehicle percentage to meet the relaxation criteria in period t PSuperM ax Constant Maximum allowance for relaxation in Super Credit System cCO2Constant Unit penalty payment cost per g/km for each vehicle sold SCAFC tVariable Total standard CAFC credit provided in period t ACAFC tVariable Actual CAFC credit consumed in period t SFC vt Variable Fuel consumption standard for vehicle project vin period t AFC vt Variable Actual fuel consumption for vehicle project vin period t WFC vVariable Weight factor for low emission vehicle project vin CAFC credit SN EV tVariable Total standard NEV credit for new energy vehicles required in period t AN EV tVariable Actual NEV credit gained for new energy vehicles in period t kvVariable Weighted NEV credit ratio factor for vehicle project v RtVariable Target ratio in period tdiscounted for NEV credit CDual Constant Monetary value for one dual credit in dual policy system Z. Shi /Junior Management Science 10(3) (2025) 748-780762 For the target NEV score calculation is represent in Equation (32). As for the target ratio Rt, it can vary from year to year. In 2025, the target ratio is approximately 20%, increasing to 40% by 2030, and eventually reaching 50% by the end of 2035. The target ratio can be adjusted each year, although there are no official rules announced. In this model, the target ratio is assumed to increase linearly and evenly each year, with specific percentages detailed in Table 7. Table 7: NEV Target Ratios for the Years 2025-2035 NEV Target Ratio (2025-2035) [%] Year Target Ratio 2025 20% 2026 24% 2027 28% 2028 32% 2029 36% 2030 40% 2031 42% 2032 44% 2033 46% 2034 48% 2035 50% The formula for the calculation of the actual NEV score is in Equation (31). The weighted factor kvfor the calculation of the actual NEV score could be hard to predict since the value is different for specific vehicle model, to simply the model, this model use the average score value and shown in table 8. Table 8: NEV Credits per Vehicle Type Weighted Factor per Type Vehicle Type Estimated NEV Credit PHEV 1.6 BEV 4 FCEV 4.8 For the monetary value CDualC redit , in Chen and He (2022) article, the average exchange cost for one credit is between 2600 −2900 RMB, so in this model it is assumed to be 330 Euro after the currency exchange. U.S. Emission Policy The detailed emission threshold values Elaw tin Equation (34) used in the US emission model are sourced from the standard released by the US Environmental Protection Agency (Register, 2023) and are summarized in Table 9. To maintain consistency and precision, the units in the policy, originally given in units of g/mile, have been converted to units of g/km, with values rounded to one decimal digit. 5.2. Policy Comparison Experiment This paper aims to employ mathematical models to explore the differences among three distinct emission policies in Europe, China, and the US, while comparing them to the baseline model with no emission policy in place. The results will be evaluated under two different demand scenarios: conservative and innovative demand. Each policy’s parameters have been calculated and estimated based on publicly available information. For each policy scenario and demand type, a comprehensive analysis will be conducted, including an examination of the vehicle initialization schedule and production quantities. Additionally, beyond the decision variables, the objective function will be thoroughly examined. This examination will encompass cost structures and average fleet emissions to gain insights into the impact of various emission policies. Table 10 summarizes the experiment’s outline. 5.3. Sensitivity and Factorial analysis for Different Emission policies In addition to comparing different emission policies, this paper will also conduct sensitivity and factor analyses for each emission policy to assess how policy parameters affect policy effectiveness and identify the parameters with the most significant impact on emission policies. For sensitivity analysis, each factor will be categorized into three levels: low, basic, and high, and tests will be conducted at these levels for analysis. In the factorial analysis, a 2-level factorial analysis will be performed, reducing each factor to two levels: high and low. A 1/2 fraction of the full factorial design method will be used to enhance the efficiency of the factorial analysis. For the analyzed outputs, due to the extensive testing required, only two types of outputs will be compared: the net total NPV (Net Present Value) and the percentage of low carbon emission vehicles. NPV is the objective function favored by car manufacturers, as they seek to maximize NPV. However, this objective may conflict with the the reduction of carbon emissions. Therefore, the percentage of low carbon emission vehicles will also be analyzed to assess the policy’s impact on social benefits from the government’s perspective and evaluate policy effectiveness. Table 11 summarizes the Design of Experiment (DOE) for the emission policy. 5.3.1. Super Credit System For the Super Credit system policy in Europe, four factors will be considered for the analysis. The first factor is the demand type, which includes two scenarios: conservative and innovative. The next factor is the threshold percentage for relaxation, which is defined as the percentage of low emission vehicles with less than 50 g/km emissions. The threshold is set in two stages, one before 2030 and one after 2030. Z. Shi /Junior Management Science 10(3) (2025) 748-780 763 Table 9: Federal Vehicle Emissions Standards Year 2025 2026 2027 2028 2029 2030 2031 2032 and later Emission standard (g/km)149 152 134 116 99 91 82 73 Emission standard (g/km)92.9 94.8 83.5 72.3 61.7 56.7 51.1 45.5 Table 10: Comparison of Different Policy Experiments Policy Comparison Experiment Policy Type Demand Output Super Credit System Con/Inno •Vehicle initialization schedule •Vehicle production quantity •Cost structure •Average fleet emission result Dual Credit System Con/Inno US Credit System Con/Inno Baseline Model (No emission policy) Con/Inno For the sensitivity analysis, the range of the threshold before 2030 is from set at level 5%, 15% and 25%, and after 2030, it is at the level of 25%, 35% and 45%. The third factor is the maximum allowed threshold percentage, which is at the level of 1%, 5% and 9%. Finally, the PHEV multiplier, which provides a base percentage for PHEV-type vehicles, is included in the analysis. It can be set to either "on" or "off" to test its effect on NPV value and the actual quantity of low carbon emission vehicles. These factors will be analyzed to understand their impact on the NPV value and the quantity of low carbon emission vehicles. 5.3.2. Dual Credit System In the China dual credit system, four factors will be considered for analysis. The first factor is the demand type, which is similar to the super credit system. The remaining three factors are tested by percentage changes, and they include one important indicator for calculating the CAFC score, one indicator for the NEV score, and one factor considering the exchange cost for credit score realization. The level of change for these factors is the same which are -50%, 0% and 50%. These factors will be analyzed to understand their impact on the NPV value and the quantity of low carbon emission vehicles. 5.3.3. US Credit System For the US credit system, apart from the demand type factor, one more factor which is the percentage change of regulated threshold are considered at level of -10%, 0% and 10%. Also strict or non-strict compliance would be tested as another factor in this scenario. For strict compliance meaning that the threshold could not be exceed and non-strict compliance it is assumed the penalty cost would be similar in EU with 95 = C/((g/km) ×year). 6. Result To assess the impact of various policy systems, Section 6.1 includes an analysis of the optimal production plan under four different policy scenarios. Additionally, the consideration of a multitude of parameters comes into play, with each having the potential to influence the final outcomes of these regulations. For this purpose, Section 6.2 conducts sensitivity analysis at three different levels (low, basic, and high) for selected parameters under different policies, while Section 6.3 performs a factor analysis of these parameters. 6.1. Optimal Production Plan The detailed plan includes the initialization of vehicle projects for each year from 2025 to 2035 , as well as the production quantity for four different types of powertrain technologies. In the vehicle project initialization section, the blocks are separated by powertrain technology, vehicle project size, and year. Gray blocks represent the start of a powertrain technology in that year, while white blocks indicate that the vehicle project will not be initiated. For the production quantity, each year from 2025 to 2035 is categorized by powertrain technology and filled with different patterns, as described in the stacked bar chart. Additionally, the average fleet emissions for each year were plotted on a line chart, with the EU-regulated fleet emission threshold as a reference for comparison. The objective value, composed of six main components [Capacity income at the end of the period, Capacity increase cost, Development cost, Profit, Fixed cost, Penalty cost/Dual Credit value], was depicted in a bar chart. The final Net Present Value (NPV) was shown on a line chart for further analysis. These visualizations provide insights into the different policy scenarios and their effects on production planning and emissions. Z. Shi /Junior Management Science 10(3) (2025) 748-780764 Table 11: Summary of Design of Experiments (DOE) for Emission Policy Summary for DOE of Emission Policy Policy Type Factors Level Type Outputs Super Credit System Demand Type ConInno Option NPV, EC percentage Relaxed Threshold Before 2030 [5% or 25%] After 2030 [25% or 45%]Percentage Maximum Threshold 1% or 9% Percentage PHEV Multiplier Y/N Option Dual Credit System Demand Type Con/Inno Option NPV, EC percentage % change of Standard Fuel Consumption -50% or 50% Percentage % change of exchange price -50% or 50% Percentage % change of NEV weight factor -50% or 50% Percentage US Credit System Demand Type Con/Inno Option NPV, EC percentage % change of CO2 threshold -10% or 10% Percentage Strict compliance Y/N Option Table 12: Parameter Changes for Dual Credit System Dual Credit System Parameter Changes Standard Fuel Consumption (CAFC) Exchange Price NEV Weighted Factor (NEV) Example Base (small_PHEV_low_2025) 3.6 = C330 1.6 -50% Percentage Change 1.8 = C165 0.8 +50% Percentage Change 5.4 = C495 2.4 Table 13: Parameter Changes for US Credit System US Credit System Parameter Changes Year 2025 2026 2027 2028 2029 2030 2031 2032 and later Emission standard (g/km)92.9 94.8 83.5 72.3 61.7 56.7 51.1 45.5 -10% Percentage Change 83.6 85.3 75.2 65.1 55.5 51.3 36.0 41.0 +10% Percentage Change 102.2 104.3 91.9 79.5 67.9 62.4 56.2 50.1 6.1.1. Conservative Demand The following results were calculated under the assumption of a conservative market demand for low-emission vehicles, characterized by a lower growth rate in low-emission vehicle adoption. This scenario represents a more cautious market approach towards low-emission vehicles. Detailed Portfolio Analysis Figure 3presents information for the four different emis- Z. Shi /Junior Management Science 10(3) (2025) 748-780 765 sion policies. In the scenario of no emission system, which can be considered a reference point for interpreting the other three policies, there are no penalties or restrictions on excess fleet emissions. In this scenario, only vehicle projects that contribute positively to the net present value of the portfolio are initiated. From this scenario, it is evident that all types of ICEV vehicles in each year have a positive effect on the profit margin. However, for the small-sized PHEV, FCEV, and BEV vehicles, the profit margin is consistently negative, so there is no incentive for manufacturers to produce these vehicle projects. For medium and large-sized vehicles in these powertrain technologies, as production costs are assumed to decrease due to technological development, medium and large-sized PHEV projects are initiated after 2029, mediumsized FCEV projects after 2031, large-sized FCEV projects after 2029, and all medium and large-sized BEV projects are initiated due to their positive profit margins. After analyzing the base scenario with no impact from emission policies, three different major policies can be analyzed. Figure 3shows that several vehicle types with negative profit margins are initiated in order to balance penalty costs or meet the constraints on excess fleet emissions. The Dual Credit Policy results in the lowest number, with about 70 initiated low-emission vehicle projects during the ten-year planning period. It is followed by 76 vehicle projects in the Super Credit System, while the US Credit System leads with the highest number, with about 87 initiated projects. ICEV vehicles dominate the portfolio, but in the US Credit System, due to strict emission standards after 2029, small-sized ICEV vehicles are terminated because of lower profit margins compared to medium and large-sized ICEV vehicles. For PHEV types, in the US Credit System, several small-sized PHEV projects are initiated to meet emission constraints. However, in the Super Credit System and Dual Credit System, no small-sized PHEV projects are initiated. Furthermore, at the end of 2035, in the Super Credit System and US Emission System, there is a sudden drop in the number of largesized PHEV projects, likely due to the sufficient production of zero-emission vehicles. For these projects, the profit margin decreases, potentially falling below that of zero-emission vehicle types due to increased production costs. For FCEV and BEV types of vehicles, medium and large-sized versions are initiated to balance the high CO2 emissions from ICEV vehicles, and small-sized versions are initiated as a last resort. In comparison, BEV projects are more favorable for manufacturers due to their higher profit margins. Regarding production quantities, it is evident that manufacturers tend to align with market demand and produce as many ICEV vehicles as possible. Comparing the results to the base model with no emission policy, it can be observed that all three emission policies would reduce the production quantity of ICEV vehicles, with the US Emission Policy leading to the largest reduction. In the Super Credit System, there is a sudden drop in the production of ICEV vehicles at year 2035. This drop is a result of the fleet emission threshold decreasing from 60 g/km to 45 g/km, causing manufacturers to reduce production to meet the stricter standards. For the Dual Credit Policy, the initial production quantity in 2025 is lower due to lower demand for low-emission vehicles and the lower fuel efficiency of ICEV vehicles. However, in the following years, production quantities increase since the ICEV vehicle become more fuel efficient thus turns out to become more favorable in the dual credit score system. Fleet Emission and NPV Analysis The fleet emissions and objective function composition for the four different policy scenarios are depicted in Figure 4. The red line represents the EU-regulated threshold for average CO2fleet emissions and serves as a reference to assess the impact of different emission policies. In the base model, shown by the black line, there is some decrease in fleet emissions due to technology development, but the average CO2fleet emissions consistently remain above the EU threshold meaning some regulation need to be performed as an external force to control the fleeting emission. The three different emission policies all have some effect on reducing fleet emissions in each year. In the Super Credit System (blue line), fleet emissions follow the EU-regulated threshold with some relaxation in certain periods. In the Dual Credit System (yellow line), the trend is similar to the line with no emission policy but with smaller absolute fleet emissions. However, fleet emissions are still above the EU-regulated threshold. Lastly, in the US Emission System (green line), fleet emissions strictly follow the US emission threshold, which becomes more stringent after 2028. In terms of net present value, the base scenario yields the highest number. In the Dual Credit Policy, the NPV is also high because manufacturers can earn money for producing low-emission vehicles. Under conservative demand, manufacturers would decide to produce more low-emission vehicles to earn these new vehicle credits that could be trade to earn some money. In the Super Credit System, a minor amount of penalty cost is incurred, resulting in a 12% decrease in the net present value compared to the base model. The US Emission System yields the lowest net present value, as targets must be strictly met, leading to the production of several low-profit-margin vehicle types and a 25% decrease in the objective value. 6.1.2. Innovative Demand In the Innovative demand scenario, the market is more receptive to low-emission vehicles, with a higher growth rate in the low-emission vehicle market. Detailed Portfolio Analysis In this demand scenario, the market introduction of new vehicle project is similar but with some minor differences, as described in Figure 5. In the baseline situation, mediumsized FCEV projects in 2030 would also be initiated due to higher demand, and the large-sized BEV vehicle project in 2025 would not be started to prioritize the production of ICEV types. Compared to the conservative demand scenario, Z. Shi /Junior Management Science 10(3) (2025) 748-780766 Figure 3: Optimal Portfolio for Different Policy in Conservative Demand for all three policy scenarios, manufacturers would initialize a lower quantity of low-emission vehicle types, prioritizing those with larger profit margins to balance penalties and revenues. For example, small-sized PHEV vehicles would not be initialized in the US emission system. Similarly, the smallsized FCEV project would not start in the Super Credit system Z. Shi /Junior Management Science 10(3) (2025) 748-780 767 (a) Fleet Emission (b) NPV Structure Figure 4: Policy Comparison in Conservative Demand and would have fewer years in the Dual Credit System and US emission system. The same trend also applies to the smallsized BEV vehicle projects, with fewer projects started after 2029 in the Super Credit System due to a larger demand for other, more profitable low or zero-emission vehicle projects that work to balance the CO2emissions. Comparing the three policies, the Dual Credit policy is much less sensitive to the demand scenario for vehicle project initialization because of its credit system policy. The initialized vehicle projects in the optimal project portfolios for the innovate demand are similar to the projects for the conservative demand since there is no threshold but monetary incentives for manufacturers. For the Super Credit System and US Emission System, producers choose the vehicle type with the lowest emissions and the highest profit margin to balance the extra emissions from ICEV types and avoid penalty costs. Once the threshold is met, there is no incentive for manufacturers to produce additional low-emission vehicles. However, for the Dual Credit System, it is always profitable to produce more low-emission vehicles because manufacturers can earn more money for the extra credits earned. The production quantity graph in Figure 5also shows a similar trend. In the Dual Credit System, the total production quantity of FCEV and BEV vehicles consistently increases as the production of ICEV types decreases. However, for the Super Credit System and US Emission System, the total production quantity is lower due to reduced demand for ICEV types of vehicles. Additionally, in the innovative demand scenario, the composition of production quantities for different powertrain types does not change significantly. However, in the Dual Credit Policy scenario, the production quantity of zero-emission vehicles increases to earn more money through credits. Fleet Emission and NPV Analysis In the Innovative demand scenario, the line plot and bar plot were used to analyze fleet emissions and NPV values, as shown in Figure 6. The black line represents the baseline model with no emission policy. After 2031, due to increased market demand for low emission vehicles, fleet emissions naturally fall below the EU regulated threshold. For the Dual Credit Policy system (yellow line), fleet emissions are much lower, reaching their lowest point after 2032. In the Super Credit System (blue line), emissions follow the EU regulated threshold until 2031, after which they drop further due to market demand. The US Emission System (green line) shows a similar trend, but after 2032, the emission threshold is lower than the EU regulated threshold. In the NPV structure graph, it can be observed that in the Innovative demand scenario, compared with the conservative demand situation, the differences in objective values between different policies are smaller. The Dual Credit Policy has a higher NPV value, about 4% more compared to the baseline model, due to the extra credits earned. For the Super Credit Policy and US Credit Policy, the objective function values are similar, both about 6% lower. The major reduction occurs before 2031, as after this year, the market itself becomes more favorable toward low emission vehicles, and the regulations have less or no effect on restricting manufacturers from producing more low emission vehicle types. 6.2. Sensitivity Analysis In the context of sensitivity analysis for the policy factors, this paper selects up to three key factors for each emission policy. These factors are chosen based on their presumed significance on the policy outcomes and their potential for being readily adjusted by governmental authorities. For each factor analyzed, this study employs stacked bar charts to compare the Net Present Value (NPV) structure and the production quantities of different vehicle types. Additionally, it includes objective values and the percentage of low emission vehicles in the total production as represented in the line on the bar chart. The parameters are categorized into three levels with Z. Shi /Junior Management Science 10(3) (2025) 748-780768 Figure 5: Optimal Portfolio for Different Policy in Innovative Demand same intervals: high, basic, and low. The rationale behind this categorization is twofold: first, to reduce the experimentation process time, as each instance typically takes around five minutes to yield results, and second, the factors would not influence the trends. Further details regarding the parameter adjustment range can be found in Section 5.3. Z. Shi /Junior Management Science 10(3) (2025) 748-780 769 (a) Fleet Emission (b) NPV Structure Figure 6: Policy Comparison in Innovative Demand 6.2.1. Super Credit System Under the Super Credit System, this analysis focuses on three crucial factors: the regulated percentage of low emission vehicles, the maximum allowable relaxation percentage, and the PHEV multiplier. The regulated low emission vehicle percentage represents the minimum proportion of low emission vehicles (those emitting less than 50 g/km of CO2) that must be met before relaxation of the emission threshold is permitted. The maximum allowed relaxation percentage sets the upper limit for threshold relaxation. Finally, the PHEV multiplier determines whether PHEV-type vehicles receive a multiplier effect, meaning that when their CO2 emissions reach 50 g/km, they are counted as approximately 0.3 of a production unit instead of 0. The mathematical formulation for these factors can be found in equations 19. Regulated EV Percentage In the objective value diagram, I observe that as the required Regulated Low Emission Vehicle Percentage increases, the objective value is slightly affected, resulting in a decrease in the final NPV value. This trend is similar for both conservative and innovative demand scenarios, as shown in the Objective Value graph in Figure 7. However, the differences are smaller in the innovative demand scenario. The primary factor driving this cost difference is the penalty cost. As the required percentage increases, it becomes more challenging for manufacturers to achieve the goal percentage needed to benefit from the Super Credit Policy. This results in a higher penalty cost, which negatively impacts the NPV value. In the production portfolio for different vehicle types, I observe a significant difference in the percentage of low emission vehicles produced as the threshold percentage is adjusted. When the threshold percentage is increased by about 10%, the percentage of low emission vehicles produced increases from 32.3% to 34.2%, representing a 2% increase. This change is mainly driven by the increase in the production volume of PHEV vehicles. On the other hand, when the threshold percentage is decreased, there is a slight reduction in the production of low emission vehicles, but this reduction is only about 0.4%, which is relatively small compared to the impact of increasing the threshold. Furthermore, in the innovative demand scenario, the production portfolio shows less variation and remains at a level of about 40%. This percentage level is higher than the percentage assumed for the conservative demand scenario and indicates that in the innovative demand scenario, a larger proportion of low emission vehicles is produced regardless of the threshold percentage. Maximum Allowed Relaxed Threshold When I analyze the change in the maximum relaxed threshold allowed in the Super Credit System, I observe that as the allowed percentage increases, the objective value also increases. The primary difference is most noticeable in the penalty costs. This trend is more evident in the conservative demand scenario, as in the innovative demand scenario, manufacturers have a stronger incentive to produce low emission vehicles. In conclusion, the relaxation of Super Credit thresholds has a smaller impact on the production portfolio in the innovative demand scenario. When examining the percentage of low emission vehicles produced in both innovative and conservative demand scenarios, I observe a stable trend with some slight differences. At allowed percentages of about 1% and 9%, the percentages are similar, likely due to changes in PHEV production. However, at higher or lower allowed percentages, manufacturers tend to produce more PHEV vehicles to increase their profits in conservative demand scenarios, while they produce fewer PHEV vehicles in innovative demand scenarios. The percentage of low emission vehicles is more stable in innovative demand, with only a 0.2% change compared to about 1% in conservative demand. Z. Shi /Junior Management Science 10(3) (2025) 748-780770 (a) Composition of the Objective Value (b) Composition of the Vehicle Portfolio Figure 7: Analysis for Change of Regulated EV Percentage (a) Composition of the Objective Value (b) Composition of the Vehicle Portfolio Figure 8: Analysis for Change of Maximum Relaxation Percentage PHEV Multiplier In the implementation of the super credit policy, the introduction of the PHEV multiplier was intended to incentivize manufacturers to produce more PHEV vehicles. However, the results of the sensitivity analysis suggest that the PHEV multiplier may not significantly influence the behavior of manufacturers, possibly due to the relatively low market demand for PHEV vehicle types. The analysis shows that there are no significant differences in both the NPV graph and the vehicle portfolio graph. The composition and absolute values remain largely unchanged, with the conservative demand scenario consistently having a higher NPV objective value and about 10% fewer low emission vehicles produced. 6.2.2. Dual Credit System In the Dual Credit System, calculating dual credits involves complexity, and this paper simplifies certain factors by using average values. The system comprises three main parts, with one crucial factor selected from each part. These parts encompass the standard fuel consumption in calculating the CAFC score for traditional ICEVs, the credit exchange price, and the NEV weight factor in calculating the NEV score for low-emission vehicles (PHEV, FCEV, BEV). To analyze the impact of these factors, adjustments of approximately 50% compared to the current assumptions were made. Standard Fuel Consumption (CAFC) The change in the standard fuel consumption criteria has a notable impact on the manufacturer’s objective value. With stricter restrictions, the objective value decreases because the manufacturer’s ability to earn dual credits as extra profit diminishes. Conversely, as the standard fuel consumption index becomes more relaxed, the objective value increases. Specifically, a 50% decrease in the index leads to a 27% decrease in the objective value in a conservative demand scenario and an 18% decrease in an innovative demand scenario. Conversely, when the index becomes more relaxed, the objective value increases by approximately 16% in both demand scenarios. Regarding the percentage of different vehicle types produced, in a conservative demand scenario, a stricter Standard Fuel Consumption index leads to a significant increase in the percentage of low emission vehicles produced, approximately 6%. Conversely, a more relaxed index results in a minor decrease in the percentage of low emission vehicles produced, about 0.4%. In an innovative demand scenario, Z. Shi /Junior Management Science 10(3) (2025) 748-780 771 (a) Composition of the Objective Value (b) Composition of the Vehicle Portfolio Figure 9: Analysis for Change of PHEV Mutiplier (a) Composition of the Objective Value (b) Composition of the Vehicle Portfolio Figure 10: Analysis for Change of CAFC Standard Fuel Consumption a stricter index leads to a smaller increase of around 1% in the production of low emission vehicles, compared to a conservative demand scenario. A more relaxed index in the innovative demand scenario also results in a modest increase of approximately 0.6% in the percentage of low emission vehicles produced, along with an increase in total production volume to earn more dual credit. Exchange Price The change in the exchange price for the dual credit policy affects both the objective value and the production portfolio. When the exchange price increases, the objective value for the manufacturer also increases, but the extent of the increase is smaller compared to changes in the Standard Fuel Consumption index in CAFC credit calculation. This trend and value of increment are similar in both conservative and innovative demand scenarios, with about a 50% increase in the price resulting in about a 2% to 5% increase in NPV. Regarding the production portfolio, as the price increases, manufacturers tend to produce more low emission vehicles to earn the dual credit value. The demand scenario does not significantly affect the trend, and the percentage increase tends to follow a logarithmic pattern rather than a linear one. With a larger price, there is a lower increase rate in the production of low emission vehicles. NEV Weighted Factor (NEV) The NEV weight factor index is a critical factor in determining the NEV score in the dual credit policy system. This factor determines how much low emission vehicles are counted in calculating the NEV score. A higher NEV weight factor index score gives low emission vehicles a higher score in the NEV score calculation, which can help manufacturers earn more value. As the NEV weight factor index increases, the objective value also increases. This trend is consistent in both demand scenarios and increases linearly by about 4. Regarding the percentage of the production portfolio, an increase in the NEV weight factor index motivates vehicle manufacturers to produce more low emission vehicles in both demand scenarios. The increase is also linear, with a slightly lower rate of increase in the innovative demand scenario, but it results in about a 1% increase in the absolute value in both scenarios. The increase in low emission vehicle production is more prominent for FCEV vehicles since FCEV has the highest NEV weight factor compared to other vehicle types. Z. Shi /Junior Management Science 10(3) (2025) 748-780778 (a) Main Effect Plot (b) Interaction Plot Figure 22: Factor analysis in US Credit System for NPV value (a) Main Effect Plot (b) Interaction Plot Figure 23: Factor analysis in US Credit System for EV percentage 7. Conclusion and Outlook This article employs a mathematical model, primarily a mixed-integer linear model, to characterize car manufacturers’ production portfolios under different emission policies. The aim is to gain a more quantitative understanding of various policy systems and their effectiveness. By integrating actual datasets with different demand scenarios and policy parameters, the study simulates the impact of these policies, assuming that manufacturers strive to maximize their profits. The results indicate that all three policy systems in Europe, China, and the United States contribute to increased production of low-emission vehicles compared to the base model with no policy in place. Regarding the initialization of low-emission vehicles, the US emission policy leads to the most significant increase in the number of vehicle initiations, while the Super Credit Policy in Europe and the Dual Credit Policy perform similarly in this regard. However, from a financial perspective, the Dual Credit Policy performs the best in preserving the car manufacturer’s profit, while the US emission policy has the most detrimental effect on the manufacturer’s profit. For the average fleet emissions, both the Super Credit Policy and the US emission policy effectively track the trend set by the regulated emission threshold. However, the Dual Credit Policy does not have a fixed threshold but rather follows emissions according to market trends. As the market increasingly favors low-emission vehicles, resulting in a shift towards an innovative demand scenario where consumers prefer such vehicles, the fleet emissions are lower in response to this trend. Furthermore, my experiments show that, it is evident that for the Super Credit Policy and the US Emission Policy, the demand level has a significant impact on the profit of the car manufacturer and the percentage of low-emission vehicles produced. In contrast, the Dual Credit Policy exhibits notable differences, where the demand level appears to influence primarily the percentage of low-emission vehicles produced but not the objective value. Speak to the specific parameter factors, the Super Credit Z. Shi /Junior Management Science 10(3) (2025) 748-780 779 System does not exert a substantial influence on either of the indicators, namely NPV (Net Present Value) and EV (Electric Vehicle) percentage. However, for the Dual Credit Policy and the US Emission Policy, these factors exhibit a relatively higher effect on both indicators, with the CAFC index being the most significant factor in the Dual Credit Policy. Notably, the market exchange price for the Dual Credit only appears to impact the net present value and not the percentage of lowemission vehicle production. In the context of the US emission policy, the only factor, which is the carbon fleet emission threshold, plays a significant role in determining the manufacturer’s profit and the policy’s effectiveness. There are several limitations to this study. Firstly, regarding the model solving method, the Super Credit System policy is not formulated as a linear model and cannot be readily transformed into a linear form for optimization. Instead, a heuristic method was employed to solve this model in two steps. While this approach may not guarantee an optimal solution, however, my experiments showed that the heuristic gives good quality solutions that resulting deviations are unlikely to have a significant impact. Secondly, in this study, market demand is based on simplified assumptions and is not specific to individual vehicle models. The total demand amount is assumed to be constant for each year in the planning period, and vehicle sales quantities may vary due to consumer preferences. Additionally, the vehicle type segmentation used in this study is relatively broad, categorizing vehicles based on powertrain technology, size, year, and power range. In reality, there is a much larger variety of vehicle models. To address this limitation, a more accurate demand prediction model could be integrated into the analysis to enhance result accuracy. Thirdly, some of the policy-specific parameters are based on assumptions. For instance, in the Dual Credit Policy system, certain parameters such as fuel consumption standard in CAFC score calculation or the weighted factor for actual NEV credit calculation are calculated based on detailed descriptions of specific vehicle models and are not included in the dataset due to privacy concerns. Obtaining a more comprehensive dataset may require closer collaboration with vehicle manufacturers. At last, in the real market scenario, the cooperation between the different vehicle manufacturer should also been considered. In this study, it is assumed that only one vehicle manufacturer exists and there would be any market trading activity occurs. These activities is hard to be modelled but if necessary could be added to this model for a more accurate result. References Ajanovic, A., & Haas, R. (2017). The impact of energy policies in scenarios on GHG emission reduction in passenger car mobility in the EU15. Renewable and Sustainable Energy Reviews,68, 1088–1096. https://doi.org/10.1016/j.rser.2016.02.013 An, F., Earley, R., & Green-Weiskel, L. (2011). Global overview on fuel efficiency and motor vehicle emission. http://www.icet.org.cn/engl ish/adminis/uploadfile/20113271414317751.pdf Bergek, A., & Berggren, C. (2014). The impact of environmental policy instruments on innovation: A review of energy and automotive industry studies. Ecological Economics,106, 112–123. https://doi.o rg/10.1016/j.ecolecon.2014.07.016 Chen, Z., & He, H. (2022, February). Policy Brief - California-China Climate Institute [Accessed: 2023].https://ccci.berkeley.edu/sites/defa ult/files/China_Dual_Credit_Policy_Brief_Feb2022.pdf Costa, C., Barbosa, J., Castro, H., Gonçalves, R., & Lanceros-Méndez, S. (2021). Electric vehicles: To what extent are environmentally friendly and cost effective? – comparative study by european countries. Renewable and Sustainable Energy Reviews,151, 111548. https://doi.org/10.1016/j.rser.2021.111548 Cuenca, R. M., Gaines, L. L., & Vyas, A. D. (2000). Evaluation of electric vehicle production and operating costs (tech. rep.). https://doi.org/10.2 172/764206 Dominioni, G., & Faure, M. (2022). Environmental Policy in Good and Bad Times: The Countercyclical Effects of Carbon Taxes and Cap-andTrade. Journal of Environmental Law,34(2), 269–286. https://d oi.org/10.1093/jel/eqac003 European Commission. (2023). CO2-Emissionsnormen für Personenkraftwagen und leichte Nutzfahrzeuge [Accessed: 2023].https://cl imate.ec.europa.eu/eu-action/transport/road-transport-reducin g-co2-emissions-vehicles/co2-emission-performance-standardscars-and-vans_de Hill, N., Skinner, I., Zazias, G., & Siskos, P. (2018, February). Assessing the impacts of selected options for regulating CO2 emissions [Transport & Environment Policy Research].https://climate.ec.europa .eu/system/files/2018-03/ldv_post_2020_co2_en.pdf Hu, Y., Wang, Z., & Li, X. (2020). Impact of policies on electric vehicle diffusion: An evolutionary game of small world network analysis. Journal of Cleaner Production,265, 121703. https://doi.org/10 .1016/j.jclepro.2020.121703 Huang, S., Shen, H., Chen, Y., Chen, M.-a., & Li, X.-j. (2015). Global organic carbon emissions from primary sources from 1960 to 2009. Atmospheric Environment,122, 505–512. https://doi.org/10.1016 /j.atmosenv.2015.10.017 International Council on Clean Transportation (ICCT). (2016). China CAFC and NEV Credits - International Council on Clean Transportation [Accessed: 2023].https://theicct.org/sites/default/files/publica tions/China%20CAFC%20and%20NEV%20credits_policy-updat e_vF_18052018_updated.pdf International Energy Agency (IEA). (2021). Fuel Economy in China – Analysis [Accessed: 2023].https://www.iea.org/articles/fuel-econo my-in-china Kellner, F., Lienland, B., & Utz, S. (2021). A multi-criteria decision-making approach for assembling optimal powertrain technology portfolios in low GHG emission environments. Journal of Industrial Ecology,25(6), 1412–1429. https://doi.org/10.1111/jiec.13148 Kieckhäfer, K., Volling, T., & Spengler, T. S. (2012). Supporting Strategic Product Portfolio Planning by Market Simulation. In Gabler Verlag, Wiesbaden (pp. 123–147). https://doi.org/10.1007/978-38349-3722-3_6 Kieckhäfer, K., Walther, G., Axmann, J., & Spengler, T. (2009). Integrating agent-based simulation and system dynamics to support product strategy decisions in the automotive industry. Proceedings of the 2009 Winter Simulation Conference (WSC), 1433–1443. https://d oi.org/10.1109/WSC.2009.5429295 Lipman, T. E., & Delucchi, M. A. (2006). A retail and lifecycle cost analysis of hybrid electric vehicles. Transportation Research Part D: Transport and Environment,11(2), 115–132. https://doi.org/10.1016/j.tr d.2005.10.002 Ma, H., Lou, G., Fan, T., Chan, H. K., & Chung, S. H. (2021). Conventional automotive supply chains under China’s dual-credit policy: Fuel economy, production, and coordination. Energy Policy,151, 112166. https://doi.org/10.1016/j.enpol.2021.112166 Michalek, J. J., Papalambros, P. Y., & Skerlos, S. J. (2005). A Study of Fuel Efficiency and Emission Policy Impact on Optimal Vehicle Design Decisions. Journal of Mechanical Design,126(6), 1062–1070. htt ps://doi.org/10.1115/1.1804195 Moran, D., Wood, R., Hertwich, E., Mattson, K., Rodriguez, J. F. D., Schanes, K., & Barrett, J. (2020). Quantifying the potential for consumer- Z. Shi /Junior Management Science 10(3) (2025) 748-780780 oriented policy to reduce European and foreign carbon emissions. Climate Policy,20(sup1), S28–S38. https://doi.org/10.1080/14 693062.2018.1551186 Ou, S., Lin, Z., Qi, L., Li, J., He, X., & Przesmitzki, S. (2018). The dual-credit policy: Quantifying the policy impact on plug-in electric vehicle sales and industry profits in China. Energy Policy,121, 597–610. https://doi.org/10.1016/j.enpol.2018.06.017 Parikh, A., Shah, M., & Prajapati, M. (2023). Fuelling the sustainable future: a comparative analysis between battery electrical vehicles (BEV) and fuel cell electrical vehicles (FCEV). Environmental Science and Pollution Research,30(20), 57236–57252. https://doi.org/10.10 07/s11356-023-26241-9 Register, F. (2023). 29184 Federal Register /Vol. 88, No. 87 /Friday, May 5, 2023 ... - GovInfo [Accessed: 2023].https://www.govinfo.go v/content/pkg/FR-2023-05-05/pdf/2023-07974.pdf Romejko, K., & Nakano, M. (2017). Portfolio analysis of alternative fuel vehicles considering technological advancement, energy security and policy [Cleaner production towards a sustainable transition]. Journal of Cleaner Production,142, 39–49. https://doi.org/10.1 016/j.jclepro.2016.09.029 Thiel, C., Nijs, W., Simoes, S., Schmidt, J., van Zyl, A., & Schmid, E. (2016). The impact of the EU car CO2 regulation on the energy system and the role of electro-mobility to achieve transport decarbonisation. Energy Policy,96, 153–166. https://doi.org/10.1016/j.enp ol.2016.05.043 Thies, C., Runde, C., Kieckhäfer, K., Wansart, J., & Spengler, T. (2022). Project portfolio planning under CO2 fleet emission restrictions in the automotive industry. Journal of Industrial Ecology.https: //doi.org/10.1111/jiec.13228 Tietge, U., Díaz, S., Yang, Z., & Mock, P. (2017, November). From Laboratory to Road - International Council on Clean Transportation [Accessed: 2023].https://theicct.org/wp-content/uploads/202 1/06/Lab-to-road-intl_ICCT-white-paper_06112017_vF.pdf TransportPolicy.net. (2018). China: Light-Duty: Fuel Consumption [Accessed: 2023].https://www.transportpolicy.net/standard/chin a-light-duty-nev/ Vacheva, G., & Hinov, N. (2019). An overview of the state of art of fuel cells in electric vehicles. 2019 International Conference on Creative Business for Smart and Sustainable Growth (CREBUS), 1–4. https ://doi.org/10.1109/CREBUS.2019.8840075 Wang, S., Zhao, F., Liu, Z., & Hao, H. (2018). Impacts of a super credit policy on electric vehicle penetration and compliance with China’s corporate average fuel consumption regulation. Energy,155.htt ps://doi.org/10.1016/j.energy.2018.05.042 Wang, Y., Zhao, F., Yuan, Y., Hao, H., & Liu, Z. (2018). Analysis of typical automakers’ strategies for meeting the dual-credit regulations regarding CAFC and NEVs. Automotive Innovation,1(1), 15–23. ht tps://doi.org/10.1007/s42154-018-0010-3 Yang, D.-X., Yang, L., Chen, X.-L., Wang, C., & Nie, P.-Y. (2023). Research on credit pricing mechanism in dual-credit policy: Is the government in charge or is the market in charge? Environment, Development and Sustainability,25(2), 1561–1581. https://doi.org/10.1007 /s10668-022-02108-w Yoro, S., & Al., E. (2020). CO2 emission sources, greenhouse gases, and the global warming effect. Woodhead Publishing. https://doi.org/10 .1016/B978-0-12-819657-1.00001-3 Zhu, X., Ding, L., Zhu, H., & Guo, Y. (2022). The dual-credit policy model, a production strategy decision-making algorithm and application to Chinese automakers. Neural Computing and Applications.http s://doi.org/10.1007/s00521-022-07703-6