Dynamic Analysis of the Impact of U.S. Tariff Policies on Automotive Production in Mexico

Abstract

Background: A system dynamics model was developed to assess how United States (U.S.) tariff policies impact production, exports, labor, and Gross Domestic Product (GDP) in the Mexican automotive sector. Methods: Beginning with a baseline scenario without tariffs, the model introduces increases of up to 25% in exports to the U.S. Additionally, it analyzes a tariff increase alongside variations in demand elasticity (−0.5 to −1.6) over a 24-month period. Results: The results indicate that a 25% tariff leads to a sustained decrease in production, GDP, and exports, although adjustments in employment occur with some delay. Specifically, production volume contracts by 34.9%, and exports contract by 31.3%. Furthermore, a greater absolute elasticity of demand corresponds to a more significant decrease in production. Conclusions: The findings indicate that geographic diversification is not an immediate substitute for the current export structure, as it requires significant logistical, organizational, and capital adjustments. Reducing dependence on the U.S. market would involve reconfiguring supply chain management, diversifying suppliers, and adapting distribution networks to gain greater flexibility. The sector’s vulnerability therefore lies in its high concentration in a single destination market, which restricts production adaptability in Mexico and weakens the capacity to respond efficiently to disruptions in the supply chain.

1. Introduction

Mexico has established itself as a major global center for the manufacturing and export of vehicles and automobile parts, ranking as the world’s seventh-largest producer of light vehicles. Owing to its proximity to the U.S., it is the fourth-largest manufacturer of auto parts, supplying 43.21% of the parts to that country [1,2].

Figure 1 shows the locations of the 16 assembly plants distributed across three major manufacturing clusters. (i) The northern border region, marked in red, includes Chihuahua, Coahuila, Nuevo León, Sonora, and Baja California, where operations by Ford, General Motors, Stellantis, Kia, Hyundai, and Toyota are concentrated. (ii) The Bajío corridor, marked in green, accounts for nearly 50% of the country’s production capacity. The general motors Honda, Ford, Mazda, Toyota, and Volkswagen are in Guanajuato; Nissan and Mercedes-Benz are in Aguascalientes; BMW and GM are in San Luis Potosí; and Querétaro, marked in blue, has established itself as a strategic hub for Tier 1 suppliers. (iii) The central region, marked in yellow, contains a variety of strategic assembly plants. Audi and Volkswagen operate in Puebla; Honda in Jalisco; Stellantis, Ford, and General Motors in the State of Mexico; JAC Motors in Hidalgo; and Nissan in Morelos. This configuration aims to leverage economies of scale, reduce the risks associated with geographic concentration, and provide an agile response to fluctuations in demand—identified in green for the Americas, blue for Europe, and yellow for Asia [3,4].

In 2024, the automotive sector achieved record production of 3.989 million light vehicles, with year-over-year growth of 5.6%; exports totaled 3.479 million units, up 5.4%, equivalent to 87.2% of Mexico’s domestic production. Moreover, the auto parts industry generated a value of USD $124.48 billion (with year-over-year growth of 3.5%), with exports to the U.S. totaling USD $37.4 billion, representing 42.8% of U.S. automotive imports [5].

This activity is also reflected in Mexico’s national economy: the automotive sector contributes 4.6% of GDP [6,7], supports more than 1.02 million direct jobs, and generates a multiplier effect that creates nearly 2 million indirect jobs, according to data from the National Auto Parts Industry (INA) [8].

Trade integration has been made possible by trade agreements, such as the North American Free Trade Agreement (NAFTA), which boosted Mexico’s exports by 53% from 1993–2020 [9], when its successor, the United States–Mexico–Canada Agreement (USMCA), came into effect, establishing stricter requirements for regional content (75% vs. 62.5% previously) and high-wage labor content (40–45% produced by workers earning more than $16/hour) [10].

Table 1 shows Mexico’s position in the main light vehicle markets, the volume of exports, and the tariff conditions in each market. In the Americas, the U.S. accounts for 87.2% of Mexican exports, subject to a 25% tariff in 2025; trade agreements with the rest of the continent (Canada, Brazil, Argentina, Colombia, Chile, and Peru) present largely untapped diversification opportunities, with only 3.7% of Mexico’s production currently exported to these destinations.

A similar dynamic is observed in Asia and Europe; however, both regions account for a small share of exports, with market shares of 1.7% and 3.5%, respectively, despite the existence of trade agreements that promote free trade, with tariffs ranging from 0% to 20% in Asia and 0% in Europe. Mexico has an opportunity to gradually diversify its light vehicle sales to other markets that together account for 65% of the global light vehicle demand [11].

Table 1 shows an extreme concentration of the Mexican automotive industry in the U.S. (87.2% of total exports), which coincides with the imposition of the highest tariff regime (25%) among all identified trading partners. This scenario contrasts with the underutilization of markets offering preferential access, particularly the European Union (3.5% of exports with 0% tariffs under the EU–Mexico Free Trade Agreement), Asia-Pacific (1.7% combined with differentiated tariffs of 0–25% depending on the country and regulatory framework), and Latin American markets (3.7% with preferential tariffs of 0–20% under bilateral agreements). This configuration represents a logistical problem because it has made transportation costs more efficient and connected supply chains with the North American market. The arrangement has created a balance that gives up geographic diversity for operational efficiency, making it harder to take advantage of tariff benefits in other markets that together make up more than 65% of the global demand for light vehicles [11].

Table 1. Light vehicle exports from Mexico and their market access conditions.

Continent	Country	Exports from Mexico	Import Tariff (Light Vehicles)	Preferential Conditions
America	U.S.	87.2%	25%	USMCA and rules of origin applicable to regional content [1,12]
	Canada	3.7%	0%	USMCA and rules of origin applicable to regional content [13]
	Brazil		0–20%	Tariff reduced under preferential agreements [14]
	Argentina		0–5%	Reduced tariff under preferential agreements [15]
	Colombia		0–20%	Preferences under the Mexico–Colombia FTA [16]
	Chile		0%	Pacific Alliance Preferential Agreement [17]
	Peru		0%	Pacific Alliance Preferential Agreement [18]
	MERCOSUR		0% (with 40% regional content); 20% outside the quota	Preferential Trade Agreement [19]
Asia	China	1.7%	20% for cars	No specific FTA for light vehicles [20]
	Japan		0%	Mexico–Japan FTA [21]
	South Korea		20%	Mexico–Korea FTA [22]
	India		20%	No trade agreement in force [23]
	Singapore		0%	ASEAN [24]
	Malaysia		0%	ASEAN Free Trade [25]
	Thailand		0–20% (depending on CPTPP compliance)	ASEAN Free Trade [25]
Europe	European Union	3.5%	0%	EU-US FTA and rules of origin [26]
	United Kingdom		0%	Post-Brexit bilateral agreement [27]
	Switzerland		0%	Bilateral agreement with the EU [26]
	Norway		0%	Agreement with the European Economic Area [26]

Note: The percentages do not sum to 100% because the table excludes marginal export destinations (accounting for the remaining 3.9%) that are not the focus of the study.

Table 1. Light vehicle exports from Mexico and their market access conditions.

Continent	Country	Exports from Mexico	Import Tariff (Light Vehicles)	Preferential Conditions
America	U.S.	87.2%	25%	USMCA and rules of origin applicable to regional content [1,12]
	Canada	3.7%	0%	USMCA and rules of origin applicable to regional content [13]
	Brazil		0–20%	Tariff reduced under preferential agreements [14]
	Argentina		0–5%	Reduced tariff under preferential agreements [15]
	Colombia		0–20%	Preferences under the Mexico–Colombia FTA [16]
	Chile		0%	Pacific Alliance Preferential Agreement [17]
	Peru		0%	Pacific Alliance Preferential Agreement [18]
	MERCOSUR		0% (with 40% regional content); 20% outside the quota	Preferential Trade Agreement [19]
Asia	China	1.7%	20% for cars	No specific FTA for light vehicles [20]
	Japan		0%	Mexico–Japan FTA [21]
	South Korea		20%	Mexico–Korea FTA [22]
	India		20%	No trade agreement in force [23]
	Singapore		0%	ASEAN [24]
	Malaysia		0%	ASEAN Free Trade [25]
	Thailand		0–20% (depending on CPTPP compliance)	ASEAN Free Trade [25]
Europe	European Union	3.5%	0%	EU-US FTA and rules of origin [26]
	United Kingdom		0%	Post-Brexit bilateral agreement [27]
	Switzerland		0%	Bilateral agreement with the EU [26]
	Norway		0%	Agreement with the European Economic Area [26]

Note: The percentages do not sum to 100% because the table excludes marginal export destinations (accounting for the remaining 3.9%) that are not the focus of the study.

Figure 2 shows the evolution of light-vehicle production (red line), exports (green line), and employment (blue line) in Mexico from January 2005 to December 2025, as documented in the data sources described on page [5] for production and exports and on page [28] for employment. As seen, the trend has been one of sustained growth; however, significant declines linked to external shocks are observed, the first of which was the U.S. recession in 2008, which affected vehicle purchases in 2009, as the U.S. was the main destination for Mexican exports. During that period, as shown in Figure 2, production and exports fell by an average of 83% over seven consecutive months, highlighting the magnitude and persistence of the crisis’s impact on the Mexican automotive sector [5]. The second decline occurred in April and May 2020, at 99% and 92%, respectively, driven by the COVID-19 pandemic, although it did not result in changes to the workforce.

The decomposition of time series via the additive method reveals recurring seasonal patterns, with peaks and troughs. Production, exports, and employment exhibit a general upward trend, with significant seasonal fluctuations that recur annually in January, February, March, and November [5]. The largest declines are observed in December, when production and exports decrease due to seasonal factors, such as the end of the fiscal year and reduced commercial activity in the sector. According to the time series decomposition analysis, a correlation of 95.5% was identified between production and exports, and a correlation of 90.1% was identified between production and employees.

Even a small change in trade policies can incur high costs, with amplified effects on globally integrated supply chains [29]. Grossman et al. (2024) [30] empirically demonstrate that trade policies can lead to substantial changes in input sourcing and search costs, quantifying welfare losses of 0.12% of GDP in highly integrated economies. The most significant empirical episodes of the past two decades include the U.S.-China Trade War (2018–2025), where tariffs escalated from 7.4% to 25% on average, forcing the reconfiguration of supply chains for companies such as Apple, Samsonite, and Cisco Systems by relocating production to Vietnam, Thailand, and Mexico [31].

The 50% tariffs on steel and aluminum implemented by the U.S. in June 2025 under Section 232 increased the average cost of a car by approximately $5790 [32]. At the same time, after Brexit, 70% of British companies reported that their supply chain costs increased by an average of 21.7%, they faced delivery times that were 30% longer, and they started stockpiling supplies, which affected their working capital [32].

Assessing potential tariff scenarios for the U.S. presidential term beginning in 2025 requires reviewing the empirical evidence of the recent protectionist cycle. During the period of 2018–2020, the vulnerability of supply chains became evident when the US imposed tariffs on various Mexican products (valued at approximately $380 billion), notably those on steel and aluminum, in May 2019 [33]. In this context of trade friction, Mexican light vehicle exports to the U.S. market registered a significant year-over-year contraction [5].

On the other hand, a review of the literature on these protectionist policies reveals complex negative externalities. Retrospective studies indicate that imposing tariffs did not achieve the objective of stimulating domestic employment in the U.S. In contrast, the U.S. manufacturing sector suffered operational deterioration because of higher prices of intermediate inputs and international trade retaliation. This historical evidence underscores that a new wave of tariffs would generate distortions that would affect the profitability and efficiency of the entire binational supply network [34,35].

Econometric projections from the University of Pennsylvania for the new U.S. presidential term in 2025 indicate that tariffs will reduce long-term GDP by approximately 5.1% and wages by 3.9% [36]; research from the Hong Kong University of Science and Technology examined whether Trump’s trade policies—in particular, the U.S.-China trade war initiated in 2018 and resumed in 2025—fulfilled their promise of revitalizing U.S. manufacturing. The results show that while tariffs favor the creation of new businesses in the U.S., the inclusion of China’s retaliatory measures in the analysis reveals that their negative impact ultimately outweighs the positive effects of trade policy [31].

The trade regime implemented in 2025 features a complex architecture of measures. On 28 January 2025, the imposition of 25% tariffs on goods imported from Canada, Mexico, and China was announced under the International Emergency Economic Powers Act (IIEPA), with implementation scheduled for 4 February of that year. However, the tariffs targeting Mexico were granted a 30-day extension, effective 4 March of the same year. On 5 and 6 March, President Trump indefinitely exempted imports; the applicable 25% tariff would be replaced with a reciprocal 12% tariff, excluding USMCA imports at a later date [37].

On 2 April 2025, reciprocal tariffs are introduced on more than 50 trading partners and non-U.S. imports from Canada and Mexico at a rate of 10%, following the expiration of the IEEPA fentanyl tariffs, supplemented by updated tariffs on automobiles, steel, and aluminum [38]. Product-specific tariffs, effective 1 February 2025, include materials essential for vehicle manufacturing, including steel, aluminum, copper, and semiconductors—critical components in structures, electrical systems, and engines—directly increasing production costs.

Table 2. Monitoring the Tariff Policies of the Trump Administration, 2025.

Target	Date of Announcement	Implementation Date	Imports Affected	Applicable Rate	Authority
Steel and Aluminum	10 February	12 March	Ending steel exemptions: $29 billion; ending aluminum exemptions: $12 billion; expanding derivatives: $44 billion under chapters 73 & 76, plus a metals content of an additional $100 billion.	25%, increased to 50%	Section 232
Cars Auto parts	12 February	April	Cars: $153 billion, Auto Parts $279 billion	25	Section 232
Copper	25 February	Not yet implemented (Potential implementation mid-July 2025, pending investigation outcome)	Copper (mined, smelted, refined, scrap copper, and derivative products)	Speculated 25% (not confirmed)	Section 232
Semiconductors	18 February 2025 (Initial announcement)/ 1 April 2025 (Formal investigation start)	Not yet implemented	Semiconductors (including semiconductor substrates, bare wafers, legacy chips, leading-edge chips, microelectronics, semiconductor manufacturing equipment (SME), and downstream products such as smartphones, laptops, and other electronics containing semiconductors)	25	Section 232

summarizes the trade policies imposed by Section 232: legislation that allows the President to adjust imports through tariffs, quotas, or license fees when the Department of Commerce determines threats to national security, specifically regarding the tariffs applied to imports of the inputs listed in the Target column, with implementation dates and applicable rates [39].

Simultaneously, the U.S. implemented support measures for local manufacturers, such as a 3.75% tax credit on the Manufacturer Suggested Retail Price (MSRP) for imported parts (valid from 3 April 2025, to 30 April 2026), which was reduced to 2.5% in the second year (through 30 April 2027), applicable exclusively to vehicles assembled in the U.S., excluding Mexican or Canadian production, creating additional competitive asymmetry that penalizes the established cross-border integration with Mexico and Canada [39].

Owing to the inherent complexity of trade policies in the automotive sector—which involves multiple nonlinear factors such as trade policies, market fluctuations, strategic decisions by stakeholders, and the location of the target company, suppliers, and end market—it is necessary to adopt a methodology capable of capturing these complex systemic relationships. System Dynamics (SD) provides a framework for understanding and simulating how interactions between economic, logistical, and political variables can influence the automotive industry, enabling analyses of spillover effects and simulation scenarios. On the basis of the analysis of key variables and their relationships, SD enables the analysis of side effects and different scenarios, and the design of robust and sustainable strategies, contributing to decision-making in dynamic and fluctuating contexts [40]. SD offers analytical capabilities to model the behavioral complexity of policies through holistic analysis and scenario simulation for strategic decision-making.

U.S. trade policies in Mexico’s automotive sector generate risks that can affect the performance of key logistics processes in supply chains, causing disruptions in production or distribution. This impacts competitiveness, production, and distribution indicators with significant economic consequences. Therefore, this article poses the following research questions:

• How do changes in tariffs affect the volume of exports in Mexico’s automotive sector?
• What effects does the increase in tariffs have on employment and GDP growth in the Mexican automotive sector?

To answer these questions, Section 2 presents a targeted literature review on the automotive supply chain in North America, the global impact of trade policies, and the application of system dynamics in this context. Section 3 describes the methodology used, including the development of an SD simulation model with definitions of variables and equations. Section 4 validates the developed model and analyzes likely scenarios based on the literature. Finally, Section 5 offers a discussion of the results, conclusions, and future research directions.

2. State of the Art

2.1. Tariffs

Tariffs are government fiscal instruments imposed on goods involved in international trade, whether through imports or exports [41]. Their use may vary according to prevailing economic conditions. From a classical perspective, tariffs are intended to protect domestic producers from foreign competition by increasing the price of imported goods, thereby reducing their presence in the domestic market and improving the relative competitiveness of local industries [42]. However, evidence from an analysis of 151 economies over the period 1963–2014 indicates that high tariffs tend to reduce both output and productivity, increase unemployment, and aggravate economic inequality [43]. In the context of supply chains, Dong and Kouvelis showed that tariff increases can reshape global network configurations by altering sourcing strategies, facility location decisions, transshipment costs, and regionalization patterns [44]. Moreover, the effects of tariff policies are not limited to the industries directly targeted; they also generate multiplier effects across the supply chain and ultimately affect final consumers [45].

2.2. Impact of Tariffs on Supply Chains

Increased tariffs trigger multiplier effects across various industrial sectors, such as manufacturing [46], technology [47], agriculture [42], textiles [48], automotive, mining and metallurgy [49], construction, and waste management [50], leading companies to adapt strategies to mitigate them, such as agricultural policy reforms aimed at environmental sustainability, taking into account variables such as climate change, population growth, and water availability [51], a combination of subsidies and restrictions on mining to strengthen the market [50] and a ban on nickel ore exports as a government strategy to increase domestic processing, although with adverse effects on supply chain resilience [49].

These measures increase the cost of imported goods, leading to higher production costs and reduced profits for companies dependent on global supply chains, with a more pronounced impact on industries such as steel, semiconductors, and chemical products, where tariffs have reduced company valuations and had mixed effects on dependent industries [42].

Consequently, tariffs affect companies through direct price changes, impacting suppliers and customers, leading to higher prices for final products, which triggers a reconfiguration of supply chains that amplifies the original shock, causing logistical disruptions, production delays, and changes in the procurement structure [30], increasing the cost of inputs, reducing productivity, and spreading effects to suppliers and customers; the magnitude of these impacts depends on the sector and the time horizon, whereas business responses include strategies for relocation and supplier diversification, as summarized in

Table 3. Economic effects by sector and channel of impact.

Sector	Affected Channel	Main Effect		Evidence
		Short-Term	Long-Term
Manufacturing	Costs, sourcing, disruption propagation, and adaptation	Tariff shocks increase costs, trigger ripple effects, and force immediate reactive adjustments	Firms redesign supply networks toward adaptability and viability, with broader effects on trade and inflation	Conceptual evidence of ripple effects, deep uncertainty, and adaptive supply chain responses under tariff shocks [46]
Technology and semiconductors	Disruption of global supply chains and components	Declines in trade	Geographic of suppliers	More intense reactions in related products [47]
Agriculture and agribusiness	Tariffs on inputs and exports	Price fluctuations and loss of markets	Supply chain adjustments and changes in external demand	Adverse effects on exports due to tariff policies [42,52]
Textile	Rapid search for production alternatives	Diversion to other countries, relocation	Reconfiguration of supply chains and potential upgrading	Historical sectoral restructuring [48]
Automotive	Components and logistics	Disruption in just-in-time supply	Supplier restructuring and nearshoring	Supply chain reconfiguration [30,48]

.

Tariffs are also associated with significant social impacts; evidence shows that tariffs implemented during the U.S.–China trade conflict in 2018–2019 reduced long-term GDP by 0.12% in the U.S., whereas China saw a decline of approximately 34% in its sales to the U.S. and a reduction of approximately 2% in its export prices [30]; a 1% increase in tariffs is associated with increases in unemployment of approximately 0.03–0.04% between 2 and 5 years after the policy’s implementation [43]. Trade retaliation by strategic partners exacerbated the situation from 2018–2019, as evidenced by the tariffs imposed by China on U.S. automobile exports, leading to a slowdown in growth [52].

The imposition of additional tariffs by the U.S. has generated systemic uncertainty in 2025, particularly in the trade dynamics within the U.S. and Chinese supply chains [53], which are characterized by their high complexity and the frequent movement of components across borders before reaching final assembly [54].

One countermeasure to tariffs could be free trade agreements, which aim to increase trade, investment, and cooperation among signatory countries; however, the USMCA is not immune to the impact of tariffs. According to 2025 projections, tariffs on Canada and Mexico would cost U.S. households more than $1200 annually [55] and affect over $1 trillion in imports, not including the effects.

2.3. Methodologies Used in Tariff Analysis

Econometric models are a quantitative approach used in tariff analysis [56], along with ordinary least squares (OLS) regression models [57] and fixed- and random-effects models, to assess the causal impact of tariffs, although they capture only partial effects at a single link, not the equilibrium feedback loops they generate [58]. Grossman et al. [30] developed a model that incorporates firm-to-firm relationships and costly search costs, calibrating parameters to match the observed responses to U.S. tariffs in China.

Blanchard et al. [59] analyzed bilateral data on applied tariffs and value-added content for 14 countries between 1995 and 2009 to study how trade policies affect the structure and performance of supply chains. They considered direct impacts such as increases in the prices of intermediate and final goods, as well as indirect effects associated with supplier reconfiguration, trade diversion, and the transmission of shocks along production chains. Furthermore, at the macroeconomic level, economies with high tariff levels experience a decline in productivity and output, accompanied by higher rates of unemployment and inequality.

Zhang et al. [60] use Autoregressive Integrated Moving Average (ARIMA) time series analysis combined with data mining (unsupervised clustering) and econometric logistic regression models, utilizing the FactSet Shipping database with information on 222 manufacturing companies for the 2018–2019 period; the analysis indicates that companies tend to increase their procurement volumes both in the lead-up to and during the implementation of tariffs. The study also identified firm size, growth potential, and profitability as key determinants of the ability to adapt to tariff disruptions. However, tariff dynamics are influenced by political factors and abrupt changes, which prevent the identification of seasonal patterns and make it difficult to generalize the results.

Rogers et al. [42] investigated 2018 tariffs, which, although designed to protect industries such as steel and electronics, had unintended consequences for supply chains, affecting both the protected firms and their suppliers and customers. The side effects of these measures suggest that protectionist policies do not always produce the expected results, highlighting the need for more in-depth analysis of the effects of tariffs on supply chains, extending beyond the industries directly affected.

Table 4. Comparative matrix of methods: data inputs, findings, and limitations.

Method	Data Required	Findings	Limitations
A calibrated structural model of global supply chains includes supplier search, negotiation, and supply reconfiguration in response to unanticipated tariffs.	Trade/manufacturing, demand elasticities, productivity and negotiation parameters, search costs, and empirical responses of prices and imports to U.S.-China tariffs.	Trump’s tariffs disrupted global supply chains through supplier renegotiation and relocation; the model estimates a welfare loss for the U.S. of 0.12% of GDP, with significant contributions from changes in sourcing and search costs [30].	This structural and calibrated approach relies on parametric assumptions and simplifications about wages, final trade, and alternative sourcing destinations; it does not constitute a direct causal estimate.
Empirical analysis using difference-in-differences and firm/product-level trade data.	Product-level import data, firm-level sourcing data, trade flows by origin country, and information on regional value content requirements under USMCA.	Stricter regional value content requirements induced nearshoring, especially toward Mexico and Canada; the effect was stronger for simpler complementary components and was associated with lower inventories, higher inventory turnover, higher average cost, and lower gross margins [61].	Focused on the U.S. automotive industry and the USMCA context; results may have limited generalizability to other sectors, countries, or trade policy regimes.
Event analysis and random effects regression to assess the impact of the 2018 tariffs on protected firms and their partners in the supply chain.	Stock prices, dates of tariff announcements/implementation, sectoral classification of firms, position in the chain, input-output tables, and financial variables.	Tariffs did not protect target firms; they generated negative financial effects and indirect consequences throughout the supply chain, with differentiated impacts for customers and suppliers [42].	Stock market-based approach, reduced ability to capture long-term operational effects, and supplier- customer relationships are inferred at the industry level rather than at the firm-to-firm level.
This study conducts a qualitative–comparative analysis of global supply chains by applying an analytical framework that examines trade policies, business strategies, and supply chain reconfiguration.	Historical-sectoral evidence on trade restrictions/agreements, firm decisions, and geographic-organizational changes in apparel, automotive, and electronics.	Historical-sectoral evidence on trade restrictions/agreements, firm decisions, and geographic-organizational changes in apparel, automotive, and electronics [48].	A noneconometric and noncausal approach requires more systematic data and broader comparisons, as the results are conditioned by sector-specific and temporal factors.
Game theory using Stackelberg and equal-power models, applied to a bilateral monopoly and duopolistic competition in a transnational supply chain with tariffs.	The theoretical parameters of the model, primarily market size, substitutability, tariffs, costs, prices, and quantities, do not rely on an observational empirical basis.	Tariffs reduce profits and consumer surplus; in bilateral monopolies, they harm the manufacturer more under manufacturer leadership; in competition, the relative effect depends on market size; and under manufacturer leadership in duopolies, the greatest benefits are obtained [56].	Theoretical approach with strong assumptions; it does not incorporate production uncertainty, logistical/technological disruptions, or detailed comparison with local production or other strategic decisions.
Empirical tools based on structural gravity models, combining partial equilibrium at the product level and general equilibrium at the product level. Aggregate, with PPML estimation of trade elasticities.	Bilateral trade, bilateral tariffs, gravitational variables, institutional dummies, and domestic production/consumption data for manufacturing.	Tariff changes affect bilateral trade, trade diversion, and welfare; in Armenia, the EAEU’s CET reduces welfare, while FTAs with Iran and China partially offset this [57].	The general equilibrium model has less detail, relies on manufacturing because of insufficient disaggregated data, may underestimate commodity-based economies, and faces risks of residual endogeneity.
Empirical analysis of global sourcing using multivariate time series clustering, ARIMA, and logistic regression	Data on maritime imports by firm, product, and month; number and variety of suppliers/countries/HTS codes; and firm financial variables.	The 2018–2019 tariffs generated heterogeneous responses in global sourcing; some firms temporarily increased inventories and then returned to previous levels, while others persistently reduced their sourcing. Furthermore, size, growth, and profitability help explain the probability of disruption [60].	Difficulty in determining the optimal number of clusters, use of aggregated data, lack of precise traceability of the actual origin of products, and possible undercapture of finer strategic decisions at the product or firm level.
Exploratory, with a qualitative approach.	Interviews with executives, evidence of supply/demand disruptions, information on lead times, procurement, inventories, risks, and resilience.	COVID-19, tariffs, and Brexit are accelerating a redesign of supply chains toward more resilient, lean, localized configurations guided by total cost rather than purchase/import cost [62].	Small sample size, qualitative evidence, lack of econometric validation, and limited scope for generalization
Fixed-effects panel model, combined with calculations of simple and cumulative tariffs on value added to measure their effect on total, forward, and backward participation in GVCs	Position in the chain, capital intensity, FDI, and educational attainment	Tariffs reduce participation in GVCs; furthermore, tariffs are amplified along the chain, and the negative effect is stronger in manufacturing, low-tech sectors, and developing countries [63].	The observational approach should consider sensitivity to sectoral and national heterogeneities, as well as the possible attenuation of effects caused by exemption regimes like the inward processing trade regime.

compares the methodologies used in tariff analysis, indicating the data they typically require, a representative example from the literature, and the reported limitations.

2.4. System Dynamics and Tariffs

Trade instruments do not operate as static tools but rather trigger dynamic, nonlinear effects that propagate disruptions through interconnected production networks. The tariffs of 2018–2019 and 2025 represented one of the largest and most abrupt shifts in U.S. trade policy in history, particularly when contrasted with the U.S. historical leadership role in reducing tariffs globally [64], marking the most significant trade shock since the COVID-19 pandemic and constituting a key event that has redefined the architecture of global supply chains [65].

Given that the above SD methodology has demonstrated its analytical versatility across various industrial sectors for assessing the ripple effect in interconnected supply networks, various authors have applied it to analyze the impact of tariffs on the economy and productive sectors.

Zhao et al. [66] employ a system dynamics model to examine how trade barriers and foreign reshoring policies affect China’s photovoltaic module exports in the broader context of the global energy transition. Their approach links the diffusion of photovoltaic installation demand across major world regions with dynamic changes in production capacity, trade barriers, reshoring intensity, prices, and export volumes. System dynamics is used to assess how tariffs and reshoring measures reshape export trajectories over time rather than only in static or short-term terms.

Hamed et al. [67] used SD to capture tariff-induced relocation dynamics through feedback loops between relative costs, supply reliability, and available capacities, demonstrating that selective tariffs can induce supplier switching in 20–40% of transactions; these shifts are accompanied by increases in the environmental impact of 5–15% due to less efficient sources. They noted limitations related to data availability, model simplification, and the need to integrate social factors and uncertainty management in the future to improve the model’s comprehensiveness and applicability.

Cai and Liu use system dynamics, integrated with life cycle assessment, to analyze how policy support, technological investment, fiscal incentives, and environmental regulation shape the long-term development of China’s automotive industry. The study explicitly recognizes that trade tariffs and exchange rate fluctuations influence the cost structure of components and finished vehicles, thereby affecting the industry’s international competitiveness. System dynamics is applied to represent the feedback relationships among government policy, R&D, industrial output, market expansion, and environmental performance, making it possible to evaluate how industrial and regulatory policies influence sectoral development over time [68].

Lai et al. [69] developed an SD model that incorporates inventory stocks from international orders, where tariffs act as exogenous policy variables affecting procurement costs and strategic sourcing decisions; however, this model focuses exclusively on the manufacturer’s perspective and currently fails to encompass a broader range of supplier types, market dynamics, and emerging trends in multisourcing.

Mohammadi et al. [70] identified a lack of quantitative models for policy-making in the steel industry’s production sector, so they developed a steel supply chain model that incorporates upstream capacity constraints and demonstrates how tariffs on raw materials are amplified across multiple stages, creating cascading effects that increase operating costs by 10–20% and delivery times by 5–15%.

Han et al. [71] investigated how export regulations, regional coalitions, and trade liberalization reshaped international rare earth trade networks. While the article is less explicit about system dynamics as a formal stock-and-flow methodology, it adopts a dynamic simulation perspective to analyze how trade barriers and coalition-based policies alter trade volumes, interdependence, and regionalization tendencies over time. The study uses simulation to capture the evolving structural effects of geopolitical trade policies, showing that export restrictions intensify regionalization, whereas trade liberalization and broader cooperation can support a more globalized and resilient supply network.

The reviewed literature highlights the versatility of SD in analyzing various types of policies across multiple industrial sectors, particularly tariffs. The studies provide evidence that SD can effectively capture the complex dynamics of supply chains and assess long-term policy impacts because of its ability to model nonlinear relationships, threshold effects, dynamic complexity, and emergent behaviors that characterize organizational responses to restrictive trade policies—which traditional econometric methods do not adequately capture. Feedback loops allow modeling of how small changes in tariff policy can generate large systemic effects through dynamic amplification. This approach enables policymakers to understand complex systems and predict the outcomes of different scenarios.

A research opportunity is identified regarding how the effects of tariffs on supply chains could alter the macroeconomic environment and restructure global logistics; the Mexican automotive sector is proposed as a case study to validate the developed model, although it can serve as a reference for other countries affected by tariff policies.

3. Methodology

A model was designed and validated via the System Dynamics methodology to assess the impact of changes in U.S. tariff policies on production in the Mexican automotive sector, adapting four stages of the methodology: conceptualization, formulation, verification, and model recommendations [72]. The stages are described below.

3.1. Conceptualization of the Model

Light vehicle production in Mexico is variable because of domestic market demand and exports to the U.S., Canada, Latin America, Europe, and Asia. This variability is due to the tariff policies of different countries that impact Mexico’s automotive exports and production, affecting the workforce and, consequently, the national GDP.

To conceptualize the model, a targeted literature review was conducted to identify the key variables that recur in research related to automotive supply chains, tariffs, logistics, and international trade. The selected variables are summarized in

Table 5. Conceptual framework and key variables of the Causal Loop Diagram (CLD).

Variable	Definition
Exports to the U.S.	Volume of light vehicles shipped from Mexico to the U.S. [52].
U.S. tariffs	Tariff rate imposed by the U.S. on Mexican vehicles [30].
Domestic market sales	Light vehicles sold in Mexico [30].
Exports	Total volume of light vehicles shipped abroad (Latin America, Europe, Asia) [43].
Exports to Other Markets	Total volume of light vehicles shipped abroad (U.S., Latin America, Europe, Asia) [30].
Final price	The unit price is determined by incorporating all relevant costs and adjustments related to the commercial and logistics processes [56].
Labor	Number of people directly employed in the manufacturing of light vehicles in Mexico [43].
Logistical requirements	A set of requirements regarding transportation, storage, distribution, and regulatory procedures for transporting vehicles from Mexican plants to destination markets [10].
Market Diversification	A strategy to increase market share in other markets (Asia, Europe, Latin America, Canada, and the domestic market) [73].
Mexico’s Gross Domestic Product (GDP)	The automotive sector’s contribution to Mexico’s GDP; it shrinks if production falls [30].
Transition Costs	Production, logistics, and regulatory compliance costs that reduce profit margins [44].
Dependence on the U.S.	Dependence on total exports concentrated in the U.S. market reflects exposure to trade disruptions originating from that destination [74].
Vehicle Inventory	Inventory of finished units that have not been allocated to a market [60].
Vehicle production in Mexico	Total number of light vehicles manufactured in Mexico [75].

.

Figure 3 shows the causal diagram depicting the relationships between the key variables identified in the literature (Table 5).

The loops are described below:

Loop R1: When exports to the U.S. decline, total export volume shrinks, leading to a buildup of finished vehicle inventory—that is, units not sold on the market. This sustained inventory surplus puts pressure on the system to diversify its markets. As this diversification overcomes logistical and commercial constraints, exports to other destinations increase, reducing dependence on the U.S. market. Consequently, as this dependence decreases, the system tends to stabilize in the long term (Figure 4).

Loop R2: When exports decline, finished product inventory accumulates, and in response, the industry reduces production. However, market diversification begins as an adjustment strategy, seeking new commercial destinations that allow production levels to return to normal. Once the new markets consolidate, exports increase, inventory decreases, and production stabilizes.

Loop B1: Increased market diversification increases the logistical requirements associated with exporting to new destinations, driving up transition costs linked to regulatory adaptation, commercial coordination, and distribution expenses. These costs reduce expected profitability and slow the consolidation of new destinations, acting as a balancing mechanism that decelerates diversification.

Loop B2: Market diversification raises transition costs (technical adaptations, compliance with foreign regulations, and contract reconfiguration). This reduces the net profit margin, creating an economic disincentive that slows further expansion.

Loop B3: Increased exports to alternative markets allow for the distribution of the accumulated inventory of finished vehicles. The reduction in this surplus mitigates the immediate need to increase diversification, so the initial momentum of the adjustment decreases in intensity, and the system tends to stabilize; the loop is shown in green.

Loop B4: The automotive sector adjusts its production levels on the basis of vehicle inventory signals; when inventories decline due to domestic sales and exports, production is adjusted accordingly.

Loop B5: When exports to the U.S. decline, total exports decrease, which increases vehicle inventory and reduces production in Mexico. However, when diversification strategies begin to increase exports to other destinations, dependence on the U.S. and exports associated with that destination decreases, leading to a more stable production environment in Mexico (Figure 5).

3.2. Model Formulation

Equations are proposed to assess the impact of changes in U.S. tariff policies on the Mexican automotive sector. The methodological design is divided into two phases:

• Phase 1. Construction and validation of the baseline model (January 2005–February 2025): In this stage, forecasting models are developed using historical data on production, exports, employment, and GDP. The objective of this phase is to calibrate and validate the model’s ability to replicate sectoral dynamics in an environment free of the proposed new tariff impositions, thereby establishing a solid baseline for comparison.
• Phase 2. Model calibration and simulation of tariff scenarios: For the period from March to December 2025, the baseline model is calibrated to quantify the initial sensitivity of production, labor, and GDP to the implementation of tariffs. Once the magnitude of this initial impact has been validated and the system calibrated, the scenarios are analyzed, evaluating the projected effects on the sector’s key variables for the time horizon spanning from January 2026 to December 2027.

3.2.1. Phase 1

An analysis of the performance of Mexico’s light vehicle automotive industry was conducted via the official repository of the National Institute of Statistics and Geography (INEGI), the agency responsible for generating and disseminating economic and social statistics in Mexico [5]. The analysis period spans from January 2005 to December 2025, comprising a time series of 252 months. The statistical analysis of the time series from the first dataset revealed that production in Mexico’s light vehicle automotive sector exhibits both a long-term upward trend and recurring seasonal patterns with annual periodicity, which is consistent with the triple exponential smoothing (additive) method proposed by Holt and Winters, which decomposes a time series into three sequential components: level, trend, and seasonality.

For direct labor in Mexico’s light vehicle automotive sector, class 336,110 of the Monthly Manufacturing Industry Survey (EMIM) was used, corresponding to the number of people directly employed by the economic unit manufacturing cars and trucks. The variable H000A (employees of the legal entity) was selected as the indicator of direct employment. The database is available from January 2018 through December 2025, forming a 96-month time series used for the analysis [28]. After various estimation methods were evaluated, a linear regression model with a deterministic trend and production-dependent harmonic seasonality was selected. Compared with simple stochastic models, this model exhibited a better fit (R²) and lower residual errors. The incorporation of a deterministic trend allows for the representation of inertial growth and structural transformations in the sector, whereas the harmonic components (sines and cosines) capture the annual seasonal cycles, helping to reduce statistical noise.

For the automotive sector’s GDP, shown in millions of Mexican pesos, historical data were compiled from Mexico’s System of National Accounts, accessed through INEGI’s official repository [76]. The time series analysis, covering the period from January 2005 to February 2025, identified a growth trajectory. To capture the dynamics, a nonlinear polynomial regression model was formulated, with a quadratic time trend and harmonic seasonality, providing a robust basis for GDP projection.

Light Vehicle Production

Equations (1)–(4) define the production forecasting model on the basis of the Holt–Winters approach. Equation (1) estimates the seasonally adjusted level component of light vehicle production by combining the observed value corrected for seasonality with the projected level from the previous period. This equation is used to update the baseline production level that underlies the monthly production forecast. The simulation is initialized with a production level of 135,448 light vehicle units.

$$l_t^P=0.96\left(y_t^P-S_{t-12}^P\right)+0.04(l_{t-1}^P+b_{t-1}^P)$$

(1)

where:

• $l_t^P\left(t\right)$: production level in period t.
• $y_t^P$: total observed production volume in month t.
• $S_{t-12}^P$: seasonal production index corresponding to the same month; Table 6.
• $l_{t-1}^P$: production level in the immediately preceding period.
• $b_{t-1}^P$: production trend for the immediately preceding period.

Equation (2) is part of the Holt–Winters production forecasting structure and is used to update the trend component $b_t^P$. This equation estimates the average rate of change in light-vehicle production over time, independent of seasonal variation. For model initialization, the trend was set at 2825.56 units, derived from the average slope of the historical production series during the first year of the analysis period (2005). In the simulation framework, this equation is used to capture the underlying growth trajectory of production and to support the dynamic updating of the baseline forecast.

$$b_t^P=0.06\left(l_t^P-l_{t-1}^P\right)+0.94b_{t-1}^P$$

(2)

Equation (3) updates the seasonal factor ${(S}_t^P).$ This equation estimates the monthly seasonal effect embedded in the historical production series by combining the current seasonal deviation with the corresponding seasonal effect from the same month of the previous year. The seasonal component is included in the model to retain the annual cyclical structure of light-vehicle production and to improve the stability and consistency of the baseline forecast.

$$S_t^P=0.16\left(y_t^P-l_{t-1}^P-b_{t-1}^P\right)+0.84S_{t-12}^P$$

(3)

Equation (4) is used to generate the monthly production forecast $P\left(t\right)$. This variable is estimated by integrating the level, trend, and seasonal terms in Equations (1)–(3). In the simulation framework, the resulting forecast is used as the baseline production trajectory against which tariff–shock scenarios are subsequently evaluated.

$$P\left(t\right)=l_t^P+h b_t^P+S_{t+h-12(k+1)}^P, k={\displaystyle \frac{h-1}{12}}$$

(4)

The initial seasonal indices ($S_t^P$) were calculated on the basis of historical production data and used to capture the recurrent cyclical structure of the automotive production process. In the model, these indices serve to isolate regular seasonal variation and avoid attributing tariff effects to fluctuations that are explained by normal production cycles, including capacity constraints, technical shutdowns, maintenance periods, and inventory adjustments (Table 6).

Table 6. Seasonal Index for Production.

Month	${\mathit{S}}_{\mathit{t}}^{\mathit{P}}$
1	−17,428.58
2	−8702.57
3	−1562.32
4	−20,395.60
5	1816.19
6	14,997.30
7	−29,992.64
8	23,937.65
9	10,907.12
10	33,815.92
11	21,867.11
12	−32,200.95

Table 6. Seasonal Index for Production.

Month	${\mathit{S}}_{\mathit{t}}^{\mathit{P}}$
1	−17,428.58
2	−8702.57
3	−1562.32
4	−20,395.60
5	1816.19
6	14,997.30
7	−29,992.64
8	23,937.65
9	10,907.12
10	33,815.92
11	21,867.11
12	−32,200.95

Light Vehicle Exports

Equation (5) was used to estimate the monthly light-vehicle export volume E(t) by combining the projected production level from Equation (4) with a seasonal component representing recurrent annual fluctuations.

To represent the 12-month seasonal cycle, harmonic terms (sine and cosine) were incorporated into the equation. Their inclusion was based on a Fourier transform analysis of the historical residual series, conducted using Python 3.14.1^® with the SciPy and NumPy libraries.

$$E\left(t\right)=727.04+0.702919P\left(t\right)+213.7334t-1626.53sin\left({\displaystyle \frac{2\pi m_t}{12}}\right)+403.275\mathrm{c}\mathrm{o}\mathrm{s}\left({\displaystyle \frac{2\pi m_t}{12}}\right)$$

(5)

where $m_t\in \left\{1,2\dots ,12\right\}$ is the monthly seasonal index for the period $t$, which takes values from 1–12 and repeats cyclically each year.

Direct Labor

Labor ($L\left(t\right)$) over time $t$ depends on current production, three lagged terms of production (Equation (4)), and labor from the previous period ($L{\left(t\right)}_{(t-1)}$) (Equation (6)). The production lags $\left({P\left(t\right)}_{(t-n)}\right)$ allow us to model firms’ operational response times; labor does not react solely to current production but adjusts gradually after observing the trend of the previous three months, thereby capturing the industry’s hiring challenges and training times. The term L${\left(t\right)}_{\left(t-1\right)}$ indicates that the labor level in each period depends heavily on its immediately preceding value.

$$\begin{array}{l}L\left(t\right)=1385.8454-0.0017788P\left(t\right)+0.0028174\mathrm{P}{\left(t\right)}_{(t-1)}-0.0027096\mathrm{P}{\left(t\right)}_{(t-2)}-0.0014637P{\left(t\right)}_{(t-3)}\\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}+0.9969058L{\left(t\right)}_{(t-1)}\end{array}$$

(6)

Gross Domestic Product

GDP behavior was modeled via a nonlinear relationship with $P\left(t\right)$, incorporating a negative quadratic term ($-0.000021P^2\left(t\right))$ to model declines associated with the automotive sector. Additionally, a second-degree polynomial with respect to time was included to adjust the curvature of the economy’s structural trend, which identifies changes in the long-term growth rate. Equation (7) is complemented by harmonic components (sin and cos) on the basis of the ability of the Fourier series to decompose and reconstruct complex cyclical patterns. In this model, the combination of both functions allows for the adjustment of both the amplitude and the phase of annual seasonality. This ensures that the model captures not only the existence of a 12-month cycle but also the exact timing (phase) of the seasonal peaks and troughs in GDP, achieving a significant reduction in the variance of the residuals compared with categorical variable methods.

$$\begin{array}{l}GDP\left(t\right)={\rho }_{GDP}\{600800.88+26.25P\left(t\right)-0.000021P^2\left(t\right)-15303.23t+145.23t^2-172383.2sin\left({\displaystyle \frac{2\pi m_t}{12}}\right)\\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em} -61051.96\mathrm{c}\mathrm{o}\mathrm{s}\left({\displaystyle \frac{2\pi m_t}{12}}\right)\}\end{array}$$

(7)

3.2.2. Phase 2

The second methodological phase was structured in two stages. The first stage covers the period from March to December 2025 for the calibration and validation of the model in the face of tariff imposition. During this time window, the initial elasticity of the system’s key variables was estimated—production $\left(P^{\tau }\left(t\right)\right)$, exports $\left(E^{\tau }\left(t\right)\right)$, labor $\left(L^{\tau }\left(t\right)\right)$, and GDP $\left({GDP}^{\tau }\left(t\right)\right)$—in response to the imposition of tariffs. Once the system was calibrated and validated, the impact of trade policies in the medium term was quantified via “what-if” scenario analysis, which was projected over the time horizon from January 2026 to December 2027. This analysis was structured by applying a tariff ${\tau }_n\left(t\right)$ differentiated by geographic destination $n$ (n = 1—U.S.; n = 2—Latin America; n = 3—Europe; and n = 4—Asia).

To estimate the elasticity of each variable in response to the imposition of tariffs during the period March–December 2025, the following procedure was applied:

• The total historical values with tariffs were quantified.
• The variables without tariffs were forecasted via the equations from Phase 1.
• The difference between the historical variables (Step 1) and the forecasted variables (Step 2) was calculated.
• The ratio between the difference from step 3 and the forecast from step 2 was calculated (Step 3/Step 2).
• Aggregate elasticity was estimated by dividing the percentage change (Step 4) by the applied tariff.

On the basis of the above, the elasticity coefficients were estimated; for production, the coefficient was −1.3756, reflecting a high negative sensitivity to the imposition of tariffs. In the case of labor and GDP, the estimated coefficients were −0.3634 and −0.288, respectively, with smaller magnitudes reflecting a moderate adjustment. Taken together, these results quantify the response of each variable to the tariff policy, providing an indicator for evaluating alternative scenarios.

Light Vehicle Production

Equation (8) estimates tariff-adjusted production. It is calculated by combining the baseline production forecast from Equation (4) with the production elasticity coefficient associated with tariff effects.

$$P^{\tau }\left(t\right)=P\left(t\right)\left(1{-1.3756\tau }_1\right)$$

(8)

Equation (9) models the automotive sector’s response to U.S. tariff barriers and the redistribution of trade flows toward alternative markets via a first-order differential equation, which reflects the dynamics of market share convergence toward a previously established diversification target, where the rate of change in market share n depends linearly on the gap between the current state and the target.

The penetration time parameter was incorporated, defined as the time required for the industry to reduce the gap between its initial market share and the diversification target. This parameter reflects operational inertia, logistical constraints, and certification processes, ensuring that the adjustment does not occur instantaneously but rather reflects the time required to modify original sales conditions and consolidate new export channels. Consequently, the model describes the gradual transition from the initial market structure to a diversified configuration, which allows for the assessment of the automotive sector’s resilience to external shocks, such as the imposition of tariff barriers, and provides an analytical framework for simulating alternative trade redistribution scenarios.

$${\displaystyle \frac{dM{C}_n\left(t\right)}{dt}}={\displaystyle \frac{M_n^{*}-M{C}_n\left(t\right)}{T_{m,n}}}$$

(9)

where:

• $M{C}_n\left(t\right)$: Market share or sales rate achieved in the $n$ market during period $t$. Initial conditions $(t=0:M{C}_1=83\%, M{C}_2=7\%, M{C}_3=6\% \mathrm{y} M{C}_4=4\%)$.
• $M_n^{*}\left(\mathrm{t}\right)$: Market diversification target. Initial values: $M_1^{*}=50\%, M_2^{*}=18\%, M_3^{*}=17\% \mathrm{y} M_4^{*}=15\%$.
• ${T_m}_n$: Months required for entry/exit in each market $n$. Assigned values: (${T_m}_1=24, {T_m}_2=48, {T_m}_3=48 y {T_m}_4=48$).

Equation (10) is used to determine the effective production $\left(P_{ef}\left(t\right)\right)$ depending on the entry or exit target for each market. This formulation allows the model to adjust production levels on the basis of the gap detected between the diversification target and the current market share for exports (${Gf}_n$) and the domestic market ($N_m$):

$$P_{ef}\left(t\right)=P^{\tau }\left(t\right)\left[{\displaystyle \sum _{n=1}^n}{Gf}_n+N_m\right]$$

(10)

The decision logic for each market is governed by the signs of the gap relative to the target.

• When the diversification target exceeds the current market share ($(M_n^{*}\left(\mathrm{t}\right)>M{C}_n\left(t\right) o {{(N}_s}^{*}>N_s))$, the goal is to reach the upper limit of market share, incentivizing the flow of production toward regions with growth potential, and the following functions are applied: ${Gf}_n=\mathrm{m}\mathrm{a}\mathrm{x}(P_{s,n},MCn,t)$ and $Nm=\max {{(N}_s}^{*},N_s)$.
• In markets where the goal is to reduce exposure, $(M_n^{*}\left(\mathrm{t}\right)<M{C}_n\left(t\right) o {{(N}_s}^{*}<N_s)$, the functions ${Gf}_n=\min \left(P_{s,n},MCn,t\right)$ and $Nm=\min {{(N}_s}^{*},N_s)$ are used. This approach limits production to the minimum available value, avoiding excess inventory and adjusting production in response to unfavorable market conditions.

For the domestic market, $N_s$ is defined as the initial market share (based on the sector’s historical performance, an initial value of 0.14 is set), and ${N_s}^{*}$ is the target domestic market share defined for market diversification.

All market share rates within the model are normalized (they sum to 1.0), ensuring consistency in resource allocation.

Light Vehicle Exports

To determine the final export volume, the impact of tariffs on exports is calculated ($E^{\tau }\left(t\right))$. In Equation (11), $\left(1{-0.3634\tau }_1\right)$ acts as a contraction coefficient, where ${\tau }_n$ represents the tariff rate.

$$E^{\tau }\left(t\right)=E\left(t\right)\left(1{-0.3634\tau }_n\right)$$

(11)

Once the export volume has been adjusted, the flow of actual exports to each market ${(E}_{ef,n}\left(t\right))$ is modeled via Equation (12) on the basis of each market’s share over time and actual production.

$${\displaystyle \frac{d{E}_{ef,n}\left(t\right)}{dt}}=P_{ef}(t)M{C}_n(t)$$

(12)

This formulation ensures that the flow to each region depends on the sector’s actual capacity for penetration and consolidation at that destination.

Direct Labor

Equation (13) projects the level of direct labor as an adjustment in staffing that responds to changes in production volume $\left(P^{\tau }\right)$. Instead of assuming an immediate adjustment, the model recognizes that the automotive sector has a staffing adjustment lag; it does not lay off or hire staff at the same rate as production changes monthly. This delay is modeled by including two- and three-month lags $(t-2) y (t-3)$, which represent the response time.

$$L^{\tau }\left(t\right)=\left[-2006.1772-0.00692666P^{\tau }\left(t\right)+{0.02142564P}^{\tau }{\left(t\right)}_{\left(t-2\right)}-{0.00925005P}^{\tau }{\left(t\right)}_{(t-3)}\right]{\displaystyle \frac{{\tau }_n}{0.25}}$$

(13)

The labor requirement in period $t$, denoted by ${\mathrm{R}\mathrm{M}\mathrm{O}}_{\mathrm{t}}$, is calculated on the basis of the ratio of employees per car ($IEA=0.232)$ and the average total export volume over the three preceding periods (Equation (14)).

$${\mathrm{R}\mathrm{M}\mathrm{O}}_{\mathrm{t}}=IEA{\displaystyle \frac{1}{3}}{\displaystyle \sum _{k=1}^3}{E}_{ef}(t-k)$$

(14)

The effective labor force $\left(L_{ef}t\right)$ is calculated via Equation (15) on the basis of the labor requirement (Equation (14)), with an initial value of $L_{ef}\left(0\right)=\mathrm{70,812}$ employees and a labor adjustment period of $T_{Ladj}$, measured in months.

To model the effective labor force $\left(L_{ef}t\right)$, Equation (15) describes the sector’s labor adjustment toward the desired level. Since changes do not occur instantaneously, the model seeks to gradually reduce the gap between required and available personnel at a rate determined by the adjustment time $\left(T_{Ladj}\right)$. This parameter reflects the administrative inertia and operational delays inherent in the hiring and layoff processes in the automotive sector, which are estimated at 2 months.

$${\displaystyle \frac{d{L}_{ef}\left(t\right)}{dt}}={\displaystyle \frac{{\mathrm{R}\mathrm{M}\mathrm{O}}_{\mathrm{t}}-{L_{ef}}_{(t-1)}}{T_{Ladj}}}$$

(15)

Gross Domestic Product

Equation (16) models the economic impact of trade restrictions on GDP. An elasticity coefficient of −0.288 is used, which quantifies the sensitivity of sectoral GDP to changes in tariff policy.

$${GDP}^{\tau }\left(t\right)=GDP\left(t\right)\left(1{-0.2887\tau }_n\right)$$

(16)

Equation (17) is the GDP response equation and estimates the effective GDP contribution of the automotive sector, denoted as ${GDP}_{ef}$. It is calculated as a function of actual production volume, $P_{ef}t$, multiplied by the marginal GDP contribution per unit produced, represented by $A_{GDP}$. In the model, this coefficient was set at 4.6% on the basis of the documented relationship between sectoral production and GDP contribution. Within the simulation framework, Equation (17) is used to translate changes in actual production into changes in the sector’s effective GDP contribution under tariff conditions.

$${\displaystyle \frac{d{GDP}_{ef}\left(t\right)}{dt}}=A_{GDP}{P}_{ef}t(t)$$

(17)

4. Results

This section presents and analyzes the results of the simulation model, systematically evaluating the impact of tariffs on production, exports, labor, and GDP for the Mexican automotive sector and scenarios of production elasticity. These scenarios should be interpreted as exploratory sensitivity exercises designed to assess how less dependence on the US market could mitigate the tariff impact under simplified assumptions.

4.1. Validation

Considering the historical data in Figure 2 and applying the equations shown in Section 3, the proposed equations are validated to generate forecast scenarios regarding the impact of tariffs on employment, production, and exports.

The statistical validation results confirm the high predictive accuracy of the proposed hybrid architecture. In the historical fit without tariffs, Pearson correlation coefficients greater than 0.90 were obtained for all variables, with particularly strong performance for employment (r = 0.976) and GDP (r = 0.958). In the simulation under tariff shocks, the model maintained significant robustness with an MAPE below 12% for Exports, GDP, and Labor. The deviation observed in the Production variable (MAPE 30.67%) is consistent with the sector’s intrinsic volatility in the face of sudden inventory adjustments; the correlation of 0.841 confirms that the trend and direction of systemic changes are correctly captured by the dynamic model (

Table 7. Statistical validation of observed historical data under scenarios with and without tariffs.

	Phase 1: Historical Data Without Tariffs				Phase 2: Historical Data with Tariffs
Variable	MAE	RMSE	MAPE (%)	Correlation Pearson (r)	MAE	RMSE	MAPE (%)	Correlation Pearson (r)
Export	22,109.85	28,456.12	8.94	0.935	31,987.56	45,618.32	11.21	0.792
Production	41,235.18	52,901.44	14.52	0.912	99,442.22	114,834.70	30.67	0.841
Labor	1421.34	1894.22	1.87	0.976	2754.11	3122.45	3.24	0.887
GDP	542,118.3	689,451.9	6.12	0.958	1,189,523.6	1,421,902.1	10.14	0.815

).

These results confirm that the model is capable of reproducing the general trends and cyclical patterns present in the historical series. For this reason, the equations in the simulation model were used to forecast the behavior of the automotive sector under different scenarios, allowing for predictions that are consistent with the observed monthly cycles and the simulation of future contexts or changes in tariff policies.

In Phase 2, Section 3.2.2, the period from March to December 2025 was used solely for the calibration of elasticity parameters and for the validation of the base model with tariffs. The scenarios designed to evaluate the imposition of tariffs and the sensitivity of demand elasticity begin in January 2026 and extend over a 24-month horizon through December 2027.

4.1.1. Scenario One: No Tariffs

Scenario one, corresponding to a free trade regime without tariffs with the U.S., serves as the baseline for examining the dynamics of the Mexican automotive sector. This scenario allows us to observe the behavior of production, exports, employment, and GDP to analyze what would have happened had tariffs not been introduced.

The production trajectory in Figure 6 shows an oscillatory pattern over the analyzed horizon, with peak expansions of 563,517 units and troughs of 206,921 units. This pattern suggests that, even under free-trade conditions, the system would operate through cycles of expansion and adjustment associated with variations in demand, inventory management, and the use of installed capacity.

The total exports shown in Figure 7 remain within a range of approximately 255,582 to 419,918 units, with periods of expansion and recovery. Even in the absence of trade restrictions, cyclical shifts in demand drive export performance. The U.S. market accounts for the majority of shipments, whereas Latin America, Europe, and Asia have a marginal share, which constitutes a vulnerability to changes in U.S. trade policy.

The labor force in Figure 8 shows a relatively stable trajectory, remaining within a range of 82,421–111,062 people, with an initial growth phase, an intermediate correction, and a subsequent partial recovery toward the end of the period. Unlike production, whose dynamics are more volatile, employment exhibits a more gradual response. This behavior indicates that adjustments in production and export activity do not translate immediately or proportionally into labor demand, revealing a partial decoupling between production and labor, with direct implications for the sector’s productivity and operational efficiency.

The GDP in Figure 9 follows a trajectory ranging from 1,722,731,148 to 4,691,589,004 million pesos, with a pattern that almost directly mirrors production dynamics, confirming the sector’s GDP dependence on its industrial performance.

The baseline scenario shows a production structure highly dependent on the U.S. market, with incipient diversification toward Latin America, Europe, and Asia and a limited capacity of the domestic market to cushion fluctuations. The strong correlation between production, employment, and GDP confirms the sector’s systemic vulnerability to external shocks.

4.1.2. Scenario Two: Tariff Increase in the U.S. 0–25%

Scenario two introduces the progressive application of tariffs in the U.S., the main destination for Mexican automotive exports. Six simulations were evaluated with 5% increases in the tariff from 0% (solid blue line), 5% (red line), 10% (pink line), 15% (green line), 20% (solid orange line), and finally 25% (purple line), to analyze the model’s sensitivity to changes in trade policy.

Figure 10 presents the monthly behavior of automotive production under different tariff levels. The top blue line corresponds to the baseline scenario, whereas the other curves reflect the trajectories as the tariff increases. Overall, all series maintain a cyclical pattern, with peaks and troughs occurring in the same periods, indicating that the system’s temporal structure remains stable. However, the scale of production decreases significantly, confirming that the impact does not alter the shape of the cycle but rather the magnitude of the manufactured volume.

In cumulative terms, the system goes from 10,075,567 units in the baseline scenario to 6,610,544 units under a 25% tariff, representing an absolute loss of 3,465,023 units and a relative contraction of −34.39%. This decline does not correspond to a marginal adjustment but rather to a structural reduction in manufacturing capacity.

Total exports fall from 8,041,894 to 5,564,260 units, with a loss of 2,477,634 and a relative change of −30.81%. Specifically, exports to the U.S. fell from 7,419,235 to 4,821,380 units, representing an absolute reduction of 2,597,855 and a contraction of 35.01%. The proximity between this decline and that of production confirms that the U.S. market serves as the primary destination for demand within the system.

In February 2026, the shock manifested when exports to the U.S. fell to 183,528 units, a monthly contraction of 28.5% compared with the baseline scenario. From that point on, the system fails to restore the previous export flow; the sector lacks the immediate capacity for trade redirection or sufficient flexibility to reconfigure its portfolio of destinations at the required speed.

The labor force confirms that the external shock not only contracts physical flows but also destroys the labor associated with the operation of the production chain (Figure 11). The cumulative workforce decreases from 2,311,556 to 1,605,716, with an absolute loss of 705,840 man-month equivalents and a relative change of −30.53%. This magnitude shows that the reduction in manufacturing activity translates into a contraction of the labor force.

Sectoral GDP fell from $83,884,621 million pesos to $54,994.32 million pesos, representing a loss of $28,890,300 million pesos and a relative change of −34.44%. The close match between this decrease and the decrease in production (−34.39%) shows a direct link between how much is made and how much economic value is created. When the U.S. market contracts, not only does production decrease, but so does the system’s capacity to transform inputs, labor, and organization into added value.

The analysis of Scenario Two reveals the high systemic vulnerability of the Mexican automotive sector to tariff increases in the U.S. Production and exports show sharp declines, with direct effects on employment and GDP. Although Latin America, Europe, and Asia increase their market share, their compensatory capacity is insufficient to offset the loss in the U.S. market, leading to further economic instability and potential job losses in the affected sectors. The negative elasticity of key variables confirms that the sector lacks structural resilience to external shocks, underscoring the urgency of designing logistics strategies to diversify markets, reduce dependence, and strengthen the domestic market. This scenario highlights the need to formulate industrial and trade policies aimed at mitigating risks and ensuring the sector’s sustainability in contexts of high global uncertainty.

4.1.3. Scenario Three: 25% Tariff and Change in Production Elasticity Coefficients

Since the U.S. is the main destination for light vehicle exports, variations in demand sensitivity to price changes directly affect production scheduling and the utilization of installed capacity in Mexico’s automotive sector. The parameters are based on the empirical evidence reported by Leard and Wu [77], who estimate an aggregate price elasticity of −0.50 for the new light vehicle market and, when disaggregating by segment, find values of −1.60 for passenger cars and −0.85 for light trucks (SUVs, pickups, and vans). The literature reference for external demand elasticity was adopted as an input for production elasticity under the assumption that the magnitude of the market reaction directly influences domestic production, as it is the main market. In the graphs, the blue line represents an elasticity of −1.60, the red line corresponds to −0.50, and the pink line reflects −1.3756, a value calculated as the coefficient of production elasticity.

The comparison reveals a systemic penalty as elasticity increases. The total output decreases from 8.63 million units in the baseline scenario to 6.47 million in the intermediate scenario and 5.91 million in the high-growth scenario, representing reductions of 25.02% and 31.43%, respectively. The same adjustment pattern is observed in total exports, sectoral GDP, and the domestic market, whereas employment shows a relatively smaller contraction in cumulative terms (−18.44% and −23.30%), suggesting a less immediate labor market response than that observed in the real variables of the system.

Figure 12 confirms that a decrease in the elasticity of external demand reduces production. The shift from an elasticity of −0.5 to −1.6 reduces the cumulative volume of production by 31.43%, whereas the intermediate scenario with an elasticity of −1.3756 shows a loss of magnitude close to that of the high scenario. In the final period, output under −1.6 represents just 68.57% of that observed with −0.5, and the −1.3756 scenario reaches 74.98%, demonstrating that lower sensitivity in the U.S. market shifts the entire trajectory toward a lower level of output.

Exports adhere to this general principle, albeit with a significantly greater spatial concentration: sales to the U.S. bear the brunt of the shock and constitute the majority of the total export contraction, whereas Latin America, Europe, and Asia experience only slight reductions. Consequently, the decline in foreign trade does not result from homogeneous redistribution among destinations but rather from high dependence on the U.S. market. This finding indicates that the export structure lacks the necessary flexibility to redirect idle capacity toward alternative markets when the U.S. becomes less elastic; therefore, greater reactivity on the part of U.S. consumers translates almost immediately into lower domestic production, lower export volumes, and lower capacity utilization.

Labor (Figure 13) behaves differently from production and exports. In the first period, Scenario −1.6 still maintains 98.2% of the employment observed in Scenario −0.5, even though production has already fallen to approximately 68.5% of the baseline. This temporary decoupling reflects the adjustment time for labor, during which it cannot be scaled back instantly at the same pace as the decline in production.

However, as the simulation horizon progresses, production gradually aligns with employment, whereas the workforce’s absorption experiences a temporary delay. During this period, the system bears a growing cost of inefficiency, as it maintains an oversized labor force relative to the new production volume. This loss of efficiency is evident in the employment-to-output ratio, which deteriorates significantly in the early periods under the lower elasticity scenario.

Sectoral GDP, on the other hand, exactly mirrors the behavior of production. The cumulative reduction in output is −25.02% for the −1.3756 scenario and −31.43% for the −1.6 scenario, identical to that observed in production and in the domestic market.

Table 8. Quantitative comparison of elasticities, scenario 3.

Variable (Cumulative/Average)	Scenario −0.5	Scenario −1.3756	Scenario −1.6	% Change (−1.3756 vs. −0.5)	% Change (−1.6 vs. −0.5)
Production	8,625,000	6,467,000	5,914,000	−25.02%	−31.43%
Exports	7,450,000	5,640,000	5,180,000	−24.29%	−30.46%
Sectoral GDP	71,800,000	53,835,000	49,233,000	−25.02%	−31.43%
Labor	2,060,000	1,680,000	1,580,000	−18.44%	−23.30%
Domestic Market	1,050,000	787,290	720,000	−25.02%	−31.43%

summarizes the cumulative and average effects of changes in the price elasticity of demand on the main variables of the automotive system. When comparing the scenarios of −0.5, −1.3756, and −1.6, it is observed that the greater absolute sensitivity of the U.S. market amplifies the contraction in production, exports, sectoral GDP, employment, and the domestic market, highlighting the system’s logistical vulnerability to changes in the behavior of the market with the greatest influence.

5. Discussion

The main results indicate that a tariff shock stops exports because other markets are not large enough or have enough capacity to take up the lost production. This situation leads to a decline in production, sectoral GDP, and domestic activity, representing a medium-term systemic risk. The following subsections provide a more detailed explanation of the different contributions:

5.1. Implications for Theory

The findings highlight the structural vulnerability of the automotive sector when production depends on a single dominant export market. Under this condition, tariff shocks cannot be absorbed efficiently because alternative destinations lack the scale and capacity required to reallocate displaced output. As a result, idle capacity accumulates, reducing production, sectoral GDP, and domestic market activity.

A tariff shock can shift an efficiency-oriented system toward a more reactive and less flexible configuration. The increase in trade costs weakens logistical fluidity, constrains supply adjustment, and intensifies the effects of demand contraction, as confirmed by scenario 3. This interpretation is consistent with previous studies showing that tariff shocks reduce supply chain flexibility and encourage reconfiguration toward regional production structures, although such transitions involve considerable adjustment costs for manufacturing networks [44].

The transition to alternative markets presents operational challenges related to just-in-time logistics, supplier coordination, assembly sequencing, and inventory management. These difficulties are reinforced by nontariff barriers such as certification requirements, safety standards, traceability rules, and local content provisions, which increase costs and reduce economies of scale. In this sense, resilience through diversification may improve risk exposure in the long run, but it can reduce efficiency in the short run, particularly in highly integrated manufacturing systems [46].

The tariff coefficients and elasticities used in the model should be interpreted as empirical approximations rather than fixed structural parameters. Their analytical purpose is to internally calibrate the direction and relative intensity of the trade shock, allowing the model to represent plausible adjustment paths within the simulated system.

5.2. Implications for Practice

For firms, competing mainly on price increases exposure to tariff shocks and external demand contraction. Automotive companies should therefore strengthen differentiation strategies on the basis of innovation, design, quality, and brand positioning to reduce demand sensitivity and improve market resilience.

Entering alternative destinations requires compliance with nontariff regulations, adaptation to different consumption patterns, and more complex coordination of suppliers, production schedules, and inventories. These adjustments tend to increase freight, compliance, and replenishment costs while also increasing the need for safety stocks and working capital.

When production declines more rapidly than labor and operational structures can adjust, unit costs rise, and competitiveness deteriorates. Firms should therefore invest in adaptive workforce planning, more agile procurement practices, and better coordination across supply networks to respond more effectively to tariff-driven disruptions.

5.3. Implications for Policy

Industrial policy should support logistical flexibility, strengthen regional integration, and preserve the cross-border coordination already established among North American suppliers and customers.

The tariffs operate as disruptive shocks that affect not only trade volumes but also production fluidity, demand responsiveness, and sectoral output. Public policy should prioritize the reduction in avoidable trade frictions through negotiated agreements, trade facilitation measures, and regulatory coordination. Such actions would help contain the broader macroeconomic costs associated with production losses and reduced domestic activity.

In addition, policies aimed at easing diversification should address nontariff barriers, improve transport and logistics infrastructure, and support firms facing the fixed costs of supplier reconfiguration. Measures that reduce adjustment asymmetries in labor and production would also contribute to preserving competitiveness in tariff-exposed industries such as automotive industries.

5.4. Limitations of the Study and Future Research

The tariff coefficients and elasticities used in the model are empirical approximations. They are useful for representing the direction and relative intensity of the trade shock, but they should not be interpreted as permanent structural parameters.

The analysis simplifies the complexity of diversification responses in global supply chains. Evolving geopolitical conditions, firm capabilities, and regulatory changes influence reconfiguration processes, potentially altering supply chain behavior over time.

Future research could extend the temporal scope of the analysis, incorporate additional exogenous variables, and examine adaptive responses such as digital coordination tools, supplier development strategies, and labor retraining mechanisms. Comparative studies across sectors and regions would also help clarify whether the vulnerability patterns identified here are specific to automotive manufacturing or reflect broader dynamics in tariff-exposed production systems.

6. Conclusions

System dynamics has proven to be a tool that allows us to isolate unobservable behaviors in purely historical data and understand the sector’s feedback loops. System dynamics has proven to be a tool that allows us to isolate unobservable behaviors in purely historical data and understand the sector’s feedback loops.

The developed model revealed that Mexico is anticipated to produce 10,075,567 light vehicle units, with 8,240,567 earmarked for export. With a 25% tariff, production volume contracts by 34.9%, and export volume decreases by 31.3%. This demonstrates that the 25% tariffs imposed by the United States generate persistent macroeconomic contractions in the Mexican automotive sector.

Although the causal diagram suggests alternative markets as a mitigation strategy, the simulation model highlights that penetration into the domestic market, Latin America, Europe, and Asia, does not eliminate risk. Market diversification involves transforming the supply chain that has historically been designed, centralized, and optimized for the U.S. Scaling up penetration in Europe, Asia, or the rest of the Americas implies rethinking the network topology: decentralizing inventory locations, developing capabilities at new port hubs, and managing longer transit times with greater variability.

The effectiveness of the market diversification strategy depends on managing technical and regulatory approvals, content requirements, and customs regulations. This involves modifying supply chain planning, supplier management with new sourcing patterns, and production with a broader product mix, which demands greater flexibility from manufacturing lines and the supply network.

In the short term, the transition to geographically dispersed demand erodes economies of scale, increasing unit logistics costs per shipment and requiring higher safety stock levels to mitigate uncertainty. Consequently, the success of this strategy should not be evaluated on the basis of export volume but rather through an analysis of the total cost of service for each market and its capacity to absorb production in a profitable and sustainable manner.

Future research should focus on investigating the time required to penetrate new markets and evaluating the actual flexibility of Mexican production lines to accommodate smaller and more heterogeneous batches. This study will examine the viability of intercontinental transport logistics from Mexico, considering the geographical proximity of competing regional production centers in Latin America, Europe, and Asia, as well as the implications of nearshoring and friendshoring strategies.