Analysis of the Relationship Between Energy Consumption in Transport, Carbon Dioxide Emissions and State Revenues: The Case of Poland

Abstract

The observed increase in demand for transport leads to higher energy consumption, which still predominantly originates from fossil fuels. As a result, the transport sector remains one of the main sources of GHG emissions. At the same time, transport activities bring economic benefits by generating public income and contributing to GDP. This article analyzes the relationship between total final energy consumption in transport, CO₂ emissions from transport, and government revenues, using Poland as a case study. The study applies Johansen’s cointegration method to identify long-term relationships between time series. Unlike many cross-country studies, this research addresses a gap by focusing specifically on Poland—a post-transition economy where the transport sector holds both fiscal and environmental significance. After cointegration test, the Impulse Response Function was used to examine dynamics and identify system shocks. Results indicate that energy consumption in transport significantly affects both CO₂ emissions and government revenues. A rise in energy use leads to higher emissions and also correlates with increased public income. The study contributes to the understanding of how transport-related energy use impacts both fiscal outcomes and emissions in a national context. These findings offer a foundation for shaping future policies that promote economic growth and stable revenue while encouraging more efficient energy use and lower emissions.

1. Introduction

Currently, a well-functioning transport sector is a fundamental driver of economic development in many countries, ensuring not only the movement of goods and passengers but also serving as a crucial component of the national energy system and financial policy [1,2]. The increasing demand for transport, driven by urbanization, globalization, and rising living standards, leads to intensified energy consumption, which still largely relies on high-emission fossil fuels [3,4]. This dominance makes the transport sector one of the main sources of greenhouse gas (GHG) emissions, including carbon dioxide (CO₂) [5], which, when released in excessive amounts, contributes to undesirable climate changes [6]. On the other hand, transport activities bring significant economic benefits, not only by generating financial resources through instruments such as excise duties and taxes but also by contributing to gross domestic product (GDP) through its close links with other economic sectors. Therefore, government policies must aim for an economic and environmental balance, ensuring stable budget revenues while simultaneously implementing mechanisms to mitigate environmental degradation [7]. This poses a challenge primarily for governments but also for policymakers and other entities responsible for shaping the economy, organizing transport, and protecting nature. The scientific community has also been consistently conducting research in this area, dedicating numerous studies to these issues.

There are many studies focused on emissions and energy consumption in transport [8,9]. Particular attention is given to the high level of emissions in road transport. An analysis of CO₂ emissions convergence from road transport in a sample of 22 European countries between 1990 and 2014 showed that these emissions tended to equalize, with the pace of this process depending on specific structural factors. Over the years, the dynamics of emission level equalization have intensified, leading to a situation where countries with lower emissions have gradually caught up with those that were previously at the forefront of pollution [10].

Given these results, the European Union’s climate goals of significantly reducing emissions in Europe by 2050 are gaining increasing importance. It is emphasized that achieving this objective requires a combined approach, including fleet electrification and an energy policy that prevents the transfer of transport emissions to other sectors [11]. Another study examined possible scenarios using Italy as a case study, estimating the demand for road transport in Italy (for the years 2005, 2019, 2022, and 2030) and assessing greenhouse gas emissions. The findings showed that even under optimistic assumptions, GHG emissions could only be reduced by 33%, which is still far below the target [12]. Therefore, to meet such ambitious reduction goals, continuous monitoring and the implementation of new strategies are necessary. However, the impact of these measures varies depending on a country’s level of economic development and transport structure [13]. In Europe, in recent years, significant progress has been made in fuel efficiency and vehicle propulsion technologies [14]. There is also potential in the digitalization of passenger transport, where technologies like Mobility-as-a-Service (MaaS) can enhance mobility while simultaneously reducing fuel consumption and air pollution [15].

The literature also highlights that fiscal policies can influence both energy consumption and carbon dioxide emissions, thereby contributing to environmental protection [16]. A study analyzing public attitudes toward changes in fossil fuel taxation in 23 European countries considered the level of political trust. The results indicated that while most Europeans (78%) believe in anthropogenic climate change and its associated risks, only 33% of respondents support tax increases [17]. Overall, changes in fuel tax systems show that raising fees often leads not only to a short-term decline in fuel consumption but also to higher tax revenue. Furthermore, it is emphasized that such additional revenues could significantly contribute to financing public transportation and developing environmentally friendly infrastructure [18]. On the other hand, it is pointed out that tax increases, despite generating higher revenues in the initial phases, may, in the long run, lead to economic deterioration, particularly in countries heavily reliant on the transport sector [19]. Another noteworthy type of government intervention involves tax incentives and subsidies for the purchase of electric vehicles, aimed at encouraging consumers to choose more environmentally friendly forms of transport. These measures are considered tools that help balance public spending with increasingly strict energy consumption and emissions requirements [20].

Publications are increasingly analyzing the impact of legislative measures related to emission reductions on national economies and the functioning of the transport sector. A key area of focus is carbon taxes imposed on fossil fuels, which are proportional to the greenhouse gas emissions generated during combustion [21]. One particularly notable case is the Carbon Border Adjustment Mechanism (CBAM), a tool designed to level the playing field and reduce “carbon leakage” outside the EU. It does so by applying import taxes (and/or export subsidies) based on the carbon content of products from countries with varying levels of climate policy stringency [22]. Another strategy analyzed by researchers is the European Union Emissions Trading System (EU ETS), which aims to gradually reduce GHG emissions in key economic sectors through the allocation and trading of CO₂ emission allowances [23]. Additionally, it is worth mentioning the continuously tightening EURO standards, which define permissible emission levels for vehicles. These regulations have led to the introduction of more environmentally friendly vehicles on European roads, improving air quality and, consequently, public health [24]. All these instruments contribute to reducing CO₂ emissions, but they may also have varied effects on the socio-economic conditions of a country. Research suggests that well-implemented climate, technological, and fiscal policies can not only effectively reduce emissions but also ensure the economic stability of a country [25,26,27].

A review of the literature also reveals that authors frequently use mathematical modeling to examine the relationships between energy consumption, emissions, and national economies. For example, the decoupling method has been applied to assess the relationship between transport-related GHG emissions and economic growth in EU countries [28]. Meanwhile, the Granger causality test has been used to analyze the relationship between energy consumption and economic growth in the ASEAN-5 countries, which include Indonesia, Malaysia, Thailand, Singapore, and the Philippines [29]. A broader approach, the Vector Error Correction Model (VECM), has been applied to evaluate the relationships between freight transport, energy consumption, economic growth, and greenhouse gas emissions in Tunisia [30]. Additionally, Johansen’s cointegration method has been used to determine causal relationships between transport infrastructure, energy consumption in transport, and economic growth [31]. Furthermore, an analogous procedure, extended with a wavelet-based likelihood ratio test, has been used to examine the long-term relationship between CO₂ emissions and real GDP in G7 countries [32].

However, less attention in the literature has been given to studying long-term relationships between energy use in transport, harmful gas emissions, and the budget revenues of the Polish state, despite the potential for analysis using appropriate statistical methods [33]. This issue is particularly relevant in the context of the Polish economy, where the transport sector plays a dual role. The relationship between government revenues and energy consumption in the transport sector is based on two-way mechanisms, well documented in the economic literature. First, fiscal policy has a direct impact on the decisions of economic entities and consumers through instruments such as excise tax on fuels, VAT, registration fees, or subsidies for low-emission means of transport. As the authors note, e.g., in [34,35], energy taxes not only play an income-generating role, but are also an effective tool for reducing negative externalities, such as pollutant emissions. It is also worth emphasizing that there is a return effect. The level of energy consumption, including transport fuels, generates revenues for the state budget, which is an important component of fiscal revenues. In countries such as Poland, where a significant part of revenues from excise tax and VAT comes from the road transport sector, variability in fuel consumption may translate into fluctuations in budget revenues.

Moreover, in the context of transition countries, fiscal policy is playing an increasingly important role in shaping the direction of modernization of the transport and energy sectors, for example through a system of tax incentives, subsidies and fiscal preferences for specific types of vehicles or technologies. As indicated by studies [36,37], such a “green fiscal transformation” can serve both environmental and budgetary purposes.

The above relationships suggest that there may be long-term relationships between fiscal revenues and energy consumption in transport, which justify the use of cointegration analysis tools. Although this study does not introduce a fully formalized theoretical model, it is based on the paradigm of mutual feedback between fiscal policy and energy decisions (so-called policy feedback loop [38]), which provides the basis for the hypothesis of the existence of a structural relationship between the variables studied.

In the empirical model, this feedback mechanism is captured through a fiscal revenue variable that reflects overall central government income, including tax revenues from energy consumption in transport. While this variable is used in aggregated form, it remains analytically meaningful in the Polish context, where revenues such as excise duties and VAT on transport fuels represent a structurally important and relatively stable component of public finances. The use of this variable reflects the functional link between transport energy consumption and budget revenues, and allows for tracing long-term interactions within the model framework.

However, while previous studies have examined similar links in broader regional or global contexts, relatively little attention has been paid to country-specific dynamics, especially in economies like Poland. As a post-transition country, Poland represents a combination of structural and institutional features that are rarely analyzed together. These include the still-important role of coal in the energy mix, the asymmetric fiscal response to price shocks in the energy sector, and the dynamic but fragmented changes in transport policy. However, in most existing studies Poland is treated as one of many points in panel or multi-country analyses, which causes specific country-level mechanisms to be aggregated or overlooked. This is why studies focusing exclusively on this country remain relevant. The study proposed in this article focuses not only on the application of the Johansen method, but also on the deliberate adjustment of the variable set to reflect the specific institutional configuration of fiscal intervention in the Polish energy sector, including environmental taxes and subsidies, which are applied differently than in Western European countries. While the method itself is not novel, its application in this particular setting is informed by the policy feedback framework, which highlights how fiscal instruments (such as fuel taxation) not only influence energy consumption and emissions, but are also shaped by them over time. This bidirectional relationship, especially visible in Poland, where a significant share of public revenue depends on fossil fuel consumption, creates a context-specific dynamic that cannot be fully captured in cross-country studies assuming more harmonized fiscal or environmental systems. Although the model used in the study does not introduce a new formal architecture, its application in the context of Poland allows for identifying long-term structural patterns between energy use, emissions, and government revenues that remain underexplored in the existing literature. In this way, the study contributes to the empirical understanding of fiscal–energy interactions in post-transition settings and may serve as a reference point for future comparative research. This research gap has therefore become the focus of this article.

The aim of this publication was to analyze the relationship between total final energy consumption in transport, carbon dioxide emissions from transport, and government revenues, using Poland as a case study. This study is based on two key research questions:

• What are the relationships between total final energy consumption in transport, CO₂ emissions from transport, and government revenues?
• What is the direction and nature of the long-term dependencies between the analyzed time series variables according to Johansen’s cointegration method?

In response to these questions, the article first presents the research background by characterizing the transport sector in Poland in terms of energy consumption, carbon dioxide emissions, and the country’s economic condition. Subsequently, the core analysis of the relationships between these variables is conducted. The study begins with an assessment of the stationarity of the examined time series. Next, Johansen’s method is applied, utilizing the Trace and Max-Eigenvalue test statistics. Once the cointegrating relationships within the analyzed time series are established, the Impulse Response Function (IRF) is calculated to gain a deeper understanding of both the short- and long-term dynamics between the variables, as well as to identify the main transmission channels (shocks in the system). The results primarily confirm the presence of long-term relationships between key variables for the transport sector, environmental sustainability, and the national economy. This is particularly important for shaping future state policies, which should aim to foster economic growth and, consequently, increase government revenues, while simultaneously optimizing energy consumption and reducing greenhouse gas emissions associated with rising transport demand. The study also has practical implications for the development of sustainable transport strategies (e.g., promoting electromobility and public transport) and for tax law reforms, ensuring stable budget revenues while limiting the use of high-emission energy sources in transport. Additionally, knowledge about shocks in the relationships between the analyzed time series can serve as a foundation for more informed energy management in the transport sector during crisis situations, such as sudden fuel price increases or the introduction of new, stringent climate regulations.

The article is structured into several sections. The study begins with an introduction to the research topic and a review of the current state of the literature related to the examined field. This section also outlines the research objective and its significance, as well as formulates the research questions. Next, a characterization of the transport sector in Poland is provided, focusing on total energy consumption, CO₂ emissions, and its connection to the country’s economic condition. Chapter three presents the materials used in the study and describes the application of Johansen’s cointegration method along with complementary analytical tools. In the following section, chapter four, the obtained results are presented and discussed. Finally, the study concludes with a general summary, formulation of key findings, discussion of limitations, and potential directions for further research.

2. The Transport Sector in Poland in Terms of Energy Use, Emissions, and Economic Condition

Poland is a country located in Central Europe, serving as a bridge between the eastern and western parts of the continent. It is a key transit hub where major transportation routes intersect, connecting regions with diverse environmental and economic conditions. As a result, Poland ranks among the most significant transit countries in Europe. Its strategic position is further reinforced by the fact that Poland’s eastern border also serves as the external border of the European Union, making it a crucial strategic and trade partner for the entire region [39]. Passenger and freight transport play a fundamental role in the economy and society, and as Poland continues to develop ongoing changes in the transport sector are observed. In this regard, legislative initiatives aimed at making transport more efficient, environmentally friendly, and safe are particularly noteworthy [40,41]. Efforts are being made to meet transport demand while simultaneously minimizing negative environmental impacts, aligning with the principles of sustainable development [42]. Another key factor driving the growing importance of transport in Poland is government budget expenditures on expanding the vehicle fleet, developing infrastructure, and enhancing the transport network. This is clearly reflected in the National Road Construction Program for 2014–2023 (with a perspective until 2025), which aims to build 3767.9 km of roads with a total budget of approximately 163.9 billion PLN.

The program includes the construction of 324.2 km of motorways, 2948.9 km of expressways, and 43 bypasses with a total length of 447.2 km [43]. Over the years, these developments have allowed citizens to travel more safely and comfortably, while ensuring that goods reach their destinations faster and without disruptions in supply chains. Expanding on this, the growing importance of modern transport solutions, including logistics process automation [44] and intelligent traffic management systems [45], is becoming increasingly evident. Innovative and environmentally friendly transport solutions, such as vehicles powered by alternative energy sources, including electricity, hydrogen or biofuels, as well as advanced emission reduction technologies, are playing an increasingly important role across the sector [46,47,48]. All these transformations are a response to contemporary economic and environmental challenges, stemming from the dependence of national development on the transport sector. Since transport serves as a critical link for multiple economic sectors, it also generates significant revenue for the state budget, primarily through fuel excise taxes, road tolls, and other transport-related taxes. Thus, economic growth and the expansion of the transport sector directly influence budget revenues. However, while transport acts as a driver of growth, it is also one of the largest consumers of energy and a major source of greenhouse gas (GHG) emissions [49,50]. The development of road infrastructure, the increasing number of vehicles, and the intensification of trade in recent decades have led to a systematic rise in energy demand within the transport sector. In Poland, transport has become the largest consumer of final energy, meaning the energy delivered to end users that excludes energy consumed by the energy sector itself [51]. After 1998, the transport sector overtook industry in energy consumption, as industrial energy use stabilized over the past few decades. Since 2018, transport has also surpassed the energy consumption of Polish households. In 2019, total transport energy consumption reached 952.6 PJ. Due to the COVID-19 crisis in 2020, it dropped to 910.5 PJ. By 2022, it had surged to a record-high 999.5 PJ [52]. These trends highlight the critical role of transport in Poland’s economy and financial stability while also pointing to challenges, particularly in terms of energy security and GHG emissions. Poland’s total energy consumption by sector from 1990 to 2022 is shown in Figure 1.

Despite the numerous efforts mentioned earlier to improve energy efficiency and implement low-emission solutions, transport in Poland, as in many other European countries, remains heavily reliant on petroleum-based products. This is primarily the result of the significant role of road transport in the overall transportation movement, as well as the widespread use of internal combustion engines as power units [53,54]. Since the 1990s, there has been an increase in total energy consumption in transport, accompanied by the continued dominance of petroleum products. Such a situation has a negative impact on pollution levels and makes the sector particularly vulnerable to fluctuations in oil prices, which in turn affects other areas of life, directly impacting consumers [55,56]. Therefore, the transition towards alternative, more eco-friendly energy sources is crucial [57]. Until 1999, aside from petroleum products, electricity was the only other significant energy source used. The energy mix became more diverse after 2007, with a noticeable but gradual rise in the share of biofuels, waste, and natural gas, indicating some attempts to diversify energy sources. However, the scale of these changes remains limited, as confirmed by data for 2022, when energy consumption in transport from petroleum products amounted to 932,145 TJ, while consumption from all other sources totaled just 67,399 TJ (Figure 2) [52].

In the context of final energy consumption in Polish transport, it is also important to note how consumption has developed in relation to different subsectors over the last few years. Passenger transport has consistently shown the highest demand for energy, which is related to the high and still increasing number of vehicles on the roads, as well as the significant role of public transport within the entire system [58,59]. Recent, noticeable disruptions in passenger transport were caused by the market collapse following the outbreak of the COVID-19 pandemic. In 2020, there was a significant decrease in the number of passengers transported, resulting from restrictions on movement and reduced societal mobility [60], which in turn led to a drop in total energy consumption. In freight transport, energy use was lower compared to passenger transport, but it also increased over time. This sector exhibited greater stability in terms of responding to economic disruptions and the crisis caused by the pandemic. Moreover, when comparing energy consumption in freight transport with that of heavy-duty vehicles, it is noticeable that in recent years, there has been an increasing demand for this type of transport. This is a result of the growing demand for goods, including due to the rise of e-commerce services, which logistics [61] must accommodate. The total final energy consumption by subsectors in the years 2020–2022 is shown in Figure 3 [52].

The increase in energy consumption, particularly from the combustion of fossil fuels, leads to higher greenhouse gas emissions, including carbon dioxide. CO₂ absorbs and retains infrared radiation, contributing to atmospheric temperature rise. When emitted excessively due to anthropogenic activities, it intensifies the greenhouse effect. In the long term, this is associated with a chain of catastrophic events, such as rising sea levels, changing climatic conditions, and altered precipitation patterns [62]. Additionally, fossil fuel combustion, which is the primary source of CO₂ emissions, also releases harmful substances, such as nitrogen oxides (NO_x) and particulate matter (PM) [63]. These pollutants degrade air quality and negatively impact human health [64]. However, a positive aspect is that, based on reports, Poland has observed a long-term downward trend in carbon oxide emissions. the most significant reduction occurred between 1990 and 2000. Despite this decline, coal remains the dominant source of CO₂ emissions, although its share in total emissions is gradually decreasing. The share of petroleum products and natural gas in CO₂ emissions has remained relatively stable over the past three decades. After 2005, there was a slight rise in emissions, likely due to the growing role of the transport sector. However, overall, fluctuations in emissions from these sources are minor, and their contribution to total emissions does not exhibit dynamic changes, unlike coal emissions. It is also important to consider data from 2019–2022, which show more noticeable fluctuations. These variations result from the COVID-19 pandemic, ongoing transitions in the energy and industrial sectors, and the increased energy demand following the crisis [65]. The CO₂ emissions by energy source in Poland between 1990 and 2022 are illustrated in Figure 4 [52].

The share of various energy sources in CO₂ emissions is reflected in the sectoral structure of these emissions. The largest source of CO₂ emissions between 1990 and 2022 was the electricity and heat production sector. However, it should be noted that this share has clearly declined year by year, mainly due to the development of green technologies, the implementation of more energy-efficient solutions, and government policies aimed at reducing heat losses and electricity consumption [66,67]. The industrial sector has also reduced emissions, albeit to a lesser extent than the energy sector. Meanwhile, emissions from the residential sector, as well as commercial and public services, have remained at a relatively similar level. Among the sectors analyzed, transport stands out with a continuous increase in emissions over the years, making it one of the few sectors that have not reduced their carbon footprint in the long term. Although economic disruptions and the COVID-19 pandemic caused temporary declines [68], the sector quickly returned to a growth trajectory. While total emissions from transport are significantly lower compared to other sectors—67.34 million tons of CO₂e in 2022 for transport, versus 140.26 million tons of CO₂e for electricity and heat production—the persistent upward trend is an undesirable phenomenon for the environment. This situation confirms that transport remains a challenge for climate policy, prompting numerous measures and strategies aimed at its decarbonization [69,70]. An example of such efforts is the European Union’s regulations on CO₂ emissions from transport. According to the revised climate targets, between 2030 and 2034, CO₂ emissions from new passenger cars in EU countries must decrease by 55%, while emissions from new vans must be reduced by 50% compared to the target levels set for 2021. Additionally, the regulations stipulate that by 2035, all new urban buses must be zero-emission [71] by 2035. Carbon dioxide emissions by sector in Poland from 1990 to 2022 are shown in Figure 5 [52].

In the context of the discussed area, it is also essential to address the economic condition of Poland. The primary indicator reflecting the level of a country’s development is Gross Domestic Product (GDP). This metric measures the total value of all goods and services produced within an economy over a specific period [72], which, in the case of transport, includes both the value of transport services and investments in infrastructure. The direct contribution of the transport sector to GDP is approximately 6%. However, its overall significance to the economy is much greater, as road transport plays a crucial role in the supply chain of sectors responsible for nearly half of Poland’s GDP [73]. Gross Domestic Product is closely linked to budget revenues, which are financial resources accumulated in public funds. In the case of transport, these revenues include proceeds from fuel excise taxes, road tolls, and taxes on business activities conducted within the transport sector [74,75]. Between 1990 and 2022, a consistent increase in both GDP and budget revenues was observed, indicating that economic growth translated into higher state budget inflows. In this context, transport plays a crucial role—government investments in infrastructure expansion and the development of the communication network support GDP growth, while tax policies influence budget revenues, thereby generating funds for further expenditures. The relationship between these indicators demonstrates that economic growth leads to a stable expansion of public finances, which, in turn, positively affects transport activity, both in freight and passenger transportation. This correlation is illustrated in Figure 6, presenting a comparison of GDP and government revenues from 1990 to 2022 [76,77].

The characteristics of Poland in terms of final energy consumption in transport, CO₂ emissions from transport activities, and state budget revenues indicate a certain relationship between these factors. Conducting further analyses using appropriate statistical methods will not only help determine the exact dependencies between these time variables but also establish the direction and nature of their long-term relationships. These findings will be presented in the following sections of the publication.

3. Materials and Methods

The article utilizes data for Poland concerning total final energy consumption in transport, carbon dioxide (CO₂) emissions from transport activities, and government revenues. The study covers the period from 1990 to 2022. Data were obtained from statistical reports by Statistics Poland (GUS) [78,79] and databases provided by the following organizations: International Energy Agency (IEA) [52], World Bank Group [77,80], and the International Monetary Fund (IMF) [76].

The considered variables were defined and expressed as follows:

• Total Final Energy Consumption in Transport (TFECT)—measured in terajoules [TJ], based on the data collection and conversion methodology of the IEA [81];
• Carbon Dioxide Emissions from Transport (CO₂T)—measured in million metric tons of carbon dioxide equivalent [Mt CO₂e], according to the World Bank approach utilizing the Emissions Database for Global Atmospheric Research (EDGAR) [82];
• Government Revenues (GREV)—expressed as a percentage of Gross Domestic Product in Poland [%GDP], based on the principles of data collection and calculations by the IMF [83]. This relative indicator enables improved comparability over time by reflecting fiscal performance relative to economic size [77].

To address potential distortions caused by currency fluctuations and inflation over the long study period, macroeconomic variables such as GDP were converted to constant 2015 U.S. dollars (USD), following the conversion methods used by the World Bank Group and the Organisation for Economic Co-operation and Development (OECD) [84]. Since GREV was expressed as a share of GDP, no further currency adjustment was required for this series. As a result, this provided a consistent and transparent basis for time series analysis.

As the variables are expressed in different units, care was taken to ensure that this dimensional heterogeneity does not bias the analysis. According to the established econometric literature [85,86] cointegration techniques are scale-invariant, provided all series are integrated of the same order. This condition was verified using three complementary tests of stationarity: the Augmented Dickey–Fuller (ADF), Phillips–Perron (PP), and Kwiatkowski–Phillips–Schmidt–Shin (KPSS) tests. In addition, numerical stability was examined through simulation-based diagnostics and impulse response analysis, which revealed no signs of model distortion or instability due to differences in measurement units.

To achieve the research objective, Johansen’s cointegration test was applied. This statistical method is commonly used to analyze long-term relationships between multiple time series variables. It is frequently employed in economics and finance to examine whether stable, long-term relationships exist between non-stationary time series that may be cointegrated [87]. Cointegration refers to a situation where the variability of two or more time series is similar in the long run. Although individual series may be non-stationary, their linear combination is stationary, indicating a stable long-term equilibrium relationship between them. This means that cointegrated series do not drift apart over time, suggesting an existing connection between them [88]. In such cases, Johansen’s cointegration test is applicable, allowing for the identification of multiple cointegration vectors and determining their number [33]. This test, which assesses the cointegration of multiple time series, is based on the vector error correction model (VECM). However, before applying it, it is necessary to establish the vector autoregression (VAR) model [89]. The general vector autoregressive model, in which each variable has a vector value and matrices are used as coefficients, takes the general form given by the equation:

$$x_t={\mu +A}_1{x}_{t-1}+A_2{x}_{t-2}+{\dots +A}_k{x}_{t-k}+u_t,$$

(1)

where:

• $\mu$—mean vector value of the time series;
• $x_t$—represents the vector of time series;
• $A_i$—autoregressive coefficient matrices;
• $u_t$—error terms.

The next step after estimating the VAR model is to convert the VAR model into its cointegration form, transforming it into the vector error correction model (VECM):

$$\Delta x_t=\mu +A{x}_{t-1}+{\Gamma }_1\Delta x_{t-1}+\cdots +{\Gamma }_k\Delta x_{t-k}+u_t,$$

(2)

where:

• $\Delta x_t=x_t-x_{t-1}$—differential operator;
• ${\Gamma }_k$—matrix for each differential delay.

The Johansen test examines the null hypothesis of no cointegration, which occurs when the matrix $A=0$ [90]. The testing of the number of cointegrating relationships utilizes two statistics: the Trace test, which examines how many cointegrating vectors exist, and the Max-Eigenvalue test, which assesses whether the number of cointegrating vectors is exactly equal to a specified value [91,92].

The Trace test evaluates the number of K linear combinations in the given time series. The test statistic for this test employs the sum of the logarithms of the eigenvalues of the cointegration matrix and is defined as follows:

$${\lambda }_{trace}\left(r\right)=-T{\sum }_{i=r+1}^{k}ln(1-{\widehat{\lambda }}_i),$$

(3)

where:

• $T$—number of observations;
• $k$—number of variables in the model;
• ${\widehat{\lambda }}_i$—$i$-th eigen value estimated from the cointegration matrix;
• $r$—the number of cointegrating vectors (null hypothesis of the test).

The null hypothesis ($H_0$) assumes that the number of cointegrating relationships is $r$, meaning there are at most $r$ cointegrating vectors, whereas the alternative hypothesis ($H_1$) assumes that the number of cointegrating relationships is greater than $r$, meaning there are more than cointegrating vectors. The test is conducted iteratively, checking successive values. If ${\lambda }_{trace}$ exceeds the critical value, $H_0$ is rejected in favor of $H_1.$ However, if ${\lambda }_{trace}$ is smaller than the critical value, $H_0$ is not rejected—indicating that the number of cointegrating relationships is not greater than $r$ [93,94].

Unlike the Trace test, the Max-Eigenvalue test examines the number of cointegrating relationships individually rather than cumulatively. It tests whether the number of cointegrating vectors is exactly $r,$ by comparing the largest eigenvalue from the cointegration matrix. The test statistic is defined as:

$${\lambda }_{\mathit{max}}\left(r,r+1\right)=-Tln(1--{\widehat{\lambda }}_{r+i})$$

(4)

where:

• ${\widehat{\lambda }}_{r+i}$− (r + 1)—the largest Eigenvalue estimated from the cointegration matrix.

The null hypothesis ($H_0$) assumes that the number of cointegrating relations is exactly $r$, while the alternative hypothesis ($H_1$) assumes that the number of cointegrating relations is $r+1$—that is, there is exactly one additional cointegrating vector. If ${\lambda }_{max}$ exceeds the critical value, $H_0$ is rejected in favor of $H_1$, so there are at least $r+1$ cointegrating vectors. If ${\lambda }_{max}$ is less than the critical value, $H_0$ is not rejected, so the number of cointegrating relationships remains at $r$ [95]. Eigenvalues are crucial as they measure the strength of the long-term equilibrium relationship. Larger eigenvalues indicate stronger cointegration relationships [33,96,97].

Additionally, after establishing the cointegration relationships, the IRF function was applied to assess the dynamics between the time series and identify shocks in the system.

Within the scope of this study, mathematical modeling was conducted using R Core software (version 4.4.2) [98], primarily with the “urca” package, which provides a set of cointegration tests for time series analysis [99].

4. Results and Discussion

The study began with a graphical representation of all the examined time series. The charts sequentially present the total final energy consumption in transport (TFECT) (Figure 7), carbon dioxide emissions from transport activity (CO₂T) (Figure 8), and government revenue (GREV) (Figure 9).

Next, the stationarity of the time series of the analyzed variables was examined using the ADF (Augmented Dickey–Fuller test), PP (Phillips–Perron test), and KPSS (Kwiatkowski–Phillips–Schmidt–Shin test). For the variables to be included in further modeling, the time series they represent must be non-stationary in their raw form and stationary after first differencing. The results of the ADF test for time series derived from raw data are presented in

Table 1. Dickey–Fuller test results for raw data.

Time Series (Variable)	Test Statistic	p-value
TFECT	−1.658	0.705
CO2T	−1.647	0.710
GREV	−2.587	0.346

, the PP test in

Table 2. Phillips–Perron test results for raw data.

Time Series (Variable)	Test Statistic	p-value
TFECT	−8.895	0.559
CO2T	−8.982	0.554
GREV	−14.236	0.211

, and the KPSS test in

Table 3. Kwiatkowski–Phillips–Schmidt–Shin test results for raw data.

Time Series (Variable)	Test Statistic	p-value
TFECT	0.889	0.01
CO2T	0.887	0.01
GREV	0.883	0.01

.

The above results obtained using three different tests indicate that all the time series represented by the raw data are non-stationary. Therefore, in the next stage, the stationarity of the series after first differencing was examined. The results of the ADF test after first differencing are shown in

Table 4. Dickey–Fuller test results after first differentiation.

Time Series (Variable)	Test Statistic	p-value
TFECT	−3.664	0.044
CO2T	−3.733	0.038
GREV	−3.756	0.037

, from the PP test in

Table 5. Phillips–Perron test results after first differentiation.

Time Series (Variable)	Test Statistic	p-value
TFECT	−22.630	0.013
CO2T	−22.426	0.015
GREV	−34.312	0.001

, and from the KPSS test in

Table 6. Kwiatkowski–Phillips–Schmidt–Shin test results after first differentiation.

Time Series (Variable)	Test Statistic	p-value
TFECT	0.140	0.1
CO2T	0.138	0.1
GREV	0.067	0.1

.

Additionally, for each of the analyzed time series, a Monte Carlo simulation with 1000 replications was conducted to estimate the distribution of the Phillips–Perron (PP) test statistic under the null hypothesis of non-stationarity (random walk type process). First, simulations were performed for all time series derived from the raw data. The results are presented in Figure 10, Figure 11, Figure 12 and Figure 13.

Then, simulations were carried out for all analyzed time series after the first differentiation. The obtained results are shown in Figure 13, Figure 14 and Figure 15, respectively.

The histograms (Figure 10, Figure 11, Figure 12, Figure 13, Figure 14 and Figure 15) illustrate the distributions of the Phillips–Perron (PP) test statistic, indicating the positions of the actual test statistics as well as the corresponding p-values—both prior to and following the first differencing. This approach facilitates an evaluation of the strength of the evidence against the null hypothesis of non-stationarity. In the case of a non-stationary series (i.e., one for which the random walk hypothesis cannot be rejected), the observed PP test statistic is expected to lie near the center of the simulated distribution. The further the observed statistic is located in the left tail of the simulated distribution, the stronger the indication that the series is stationary. Accordingly, the results of the Monte Carlo simulations reveal that, for each of the analyzed time series after first differencing, the PP test statistic falls within the left tail of the simulated distribution, thereby providing substantial evidence in favor of stationarity.

To sum up, all obtained results confirm that considered time series represented by the raw data are non-stationary, whereas after first differencing, they exhibit stationarity. Therefore, it is possible to proceed with the Johansen cointegration test, starting first with the selection of the optimal lag order (

Table 7. Selection of the optimal lag order for the Johansen test.

Criteria	Lag 1	Lag 2	Lag 3	Lag 4	Lag 5
AIC(n)	40.337	39.703	39.565	37.835	32.839
HQ(n)	40.608	40.190	40.268	38.754	33.975
SC(n)	41.312	41.458	42.100	41.150	36.935
FPE(n)	3.370 × 1017	2.002 × 1017	2.391 × 1017	8.903 × 1016	3.856 × 1015

).

Given the obtained values, it can be observed that the optimal lag order is 5, and, therefore, this value was adopted for the subsequent part of the study. Next, the number of cointegration relationships was examined using two statistics—the Trace test and the Max-Eigenvalue test. The obtained results are presented below.

• Eigenvalues (λ_i)

• 0.8877;
• 0.6988;
• 0.1823;
• 0.0268.

The results of the Trace test statistic (

Table 8. Values of Trace test statistic and critical values.

Cointegrating Vectors (r)	Test	10 pct	5 pct	1 pct
r ≤ 3	0.76	6.50	8.18	11.65
r ≤ 2	6.40	15.66	17.95	23.52
r ≤ 1	39.99	28.71	31.52	37.22
r ≤ 0	101.21	45.23	48.28	55.43

) indicate that the hypotheses for r = 0 and r ≤ 1 should be rejected, but there is no reason to reject the hypothesis r ≤ 2. Thus, it is assumed that there are two cointegrating vectors. The eigenvectors matrix (

Table 9. Eigenvectors matrix—cointegration relations.

	TFECT.L5	CO2T.L5	GREV.L5	PASS.L5
TFECT.L5	1.000	1.000	1.000	1.000
CO2T.L5	−1.545 × 104	−1.669 × 104	−1.290 × 104	−1.366 × 104
GREV.L5	−1.771 × 103	8.413 × 103	−4.145 × 103	−2.389 × 102

) presents the coefficient in each cointegration relationship. The first row of this matrix is normalized with respect to the TFECT time series, which serves as the normalizing variable. This implies that each relationship is expressed relative to this series. High values for CO₂T and GREV suggest that these series play a significant role in the cointegration relationships. Therefore, cointegration includes TFECT, CO₂T, and GREV, forming linear combinations that remain stable over the long term.

Subsequently, the loading matrix (

Table 10. Loading matrix of analyzed time series.

	TFECT.L5	CO2T.L5	GREV.L5
TFECT.L5	1.614	2.280	−0.290
CO2T.L5	1.009 × 104	1.910 × 10−4	−1.524 × 10−5
GREV.L5	3.459 × 104	−9.333 × 10−5	1.617 × 10−5

) was constructed. The analysis of this matrix reveals how each time series responds to deviations from long-term equilibrium. The largest absolute values occur for TFECT (≈1.61 and ≈2.28), indicating that this series reacts strongly to deviations from equilibrium in both cointegration relationships.

As demonstrated by the results presented above, the Johansen procedure confirmed the cointegration of the time series. Therefore, the next stage of the study involves estimating the VECM model. The estimated model includes the RLM (Restricted Linear Model) and the matrix of cointegrating vectors β, which accounts for long-term relationships between the time series. These are the linear combinations of the series that remain in equilibrium over the long run. In this analysis, two cointegration relationships and three time series are considered, meaning that the β matrix has dimensions 2 × 3. The values represented by this matrix indicate how the individual series are related to one another. High coefficients suggest strong dependencies between the series. The matrix of cointegrating vectors β including trend coefficients is shown in

Table 11. Matrix of cointegrating vectors β with attached trend coefficients.

	ect1	ect2
TFECT.L5	1.000	6.776 × 10−21
CO2T.L5	0.000	1.000
GREV.L5	−34,483.640	−2.467
TREND.L5	−26,828.310	−1.807

.

In the first cointegration vector (ect1), TFECT.L5 is assigned a coefficient equal to 1, establishing it as the normalizing variable for interpreting the long-run equilibrium. The coefficient for CO₂T.L5 is 0, indicating that this series is not part of the first relationship. For GREV.L5, there is a large negative value, which suggests a strong inverse relationship with TFECT.L5. Additionally, the value of the trend coefficient suggests that there is a long-term trend affecting TFECT.L5. Based on the specified coefficients, the first cointegration relationship (ect1) can be expressed by the following equation:

$$TFECT.L5-\mathrm{34,483.64}\times GREV.L5-\mathrm{26,828.31}\times TREND.L5=0$$

(5)

In the second relationship (ect2), CO₂T is the normalizing variable, which is why its coefficient is 1. For TFECT, the coefficient is 0 because this series is not part of the second relationship. The value of the coefficient for GREV.L5 indicates that as GREV.L5 increases, CO₂T.L5 decreases. Taking the values of the coefficients into account, the second cointegration relationship (ect2) can be expressed by the following equation:

$${CO}_{2}T.L5-2.467\times GREV.L5-1.807\times TREND.L5=0$$

(6)

In the next part of the study, after identifying the cointegrating relationships present in the considered time series, the Impulse Response Function (IRF) [100] was computed. IRF analysis allows for a deeper understanding of both short- and long-term dynamics between time series, as well as identifying the main transmission channels of shocks in the system. This analysis helps in determining how a change in one time series influences the dynamics of the entire system in subsequent periods. Particularly significant are sudden changes, or shocks. They primarily allow for assessing which time series are especially sensitive to changes occurring in others. Additionally, the IRF shows the direction of the influence—whether it is direct or opposite. Moreover, it reveals how long the shock effect lasts over time, i.e., whether the system quickly returns to equilibrium or if the shock causes long-lasting changes. The value of the IRF indicates the strength of the shock’s impact. Large responses indicate a strong influence, while small ones show more subtle reactions [101,102]. The responses to shocks in the system were analyzed using all the considered time series. Figure 16 presents the impulse response from TFECT, while

Table 12. TFECT’s response to shocks in the system.

Period	TFECT	CO2T	GREV
1	19,366.846	1.387	0.435
2	16,767.475	1.154	0.339
3	17,264.957	1.123	0.464
4	4907.827	0.269	0.117
5	2431.460	0.082	0.272
6	−2059.687	−0.144	−0.116
7	1617.437	0.147	0.063
8	2350.726	0.222	0.019
9	1967.259	0.184	0.126
10	1543.445	0.152	0.152
11	2185.142	0.182	0.210

shows the IRF values from TFECT in each period.

A shock in TFECT leads to a notable rise in the first period (≈19,367), followed by a gradual decline over the next few periods. Around period 6, the effect turns slightly negative (≈−2060), after which the system settles at a new relatively stable level. This indicates a strong but short-lived reaction, with the system reaching a new equilibrium within a few periods. Regarding the shock in CO₂T, a small initial increase is observed (≈1.387 in period 1), but this effect quickly dissipates. Generally, there is little long-term impact—after several periods, the values are low. On the other hand, analyzing the shock in GREV, a very small effect is visible—a slight surge initially, which next diminishes with some fluctuations after third period.

The impulse response from TFECT shows an immediate and short-lived rise in CO₂ emissions. This is consistent with the dominant role of fossil fuels in the transport sector, where combustion leads to direct emission effects. GREV respond with a slight delay, reaching a peak in the third period (≈0.464). This lag does not result from the excise tax collection mechanism itself, but more likely reflects the time needed for increased fuel consumption to influence broader fiscal aggregates, including VAT and business-related taxes. These revenues often react indirectly or follow periodic reporting cycles. This pattern observed supports the presence of both immediate environmental effects and more gradual fiscal adjustments.

Next, the responses to shocks in the system with CO₂T were examined. The results are illustrated in Figure 17, and

Table 13. CO₂T response to shocks in the system.

Period	TFECT	CO2T	GREV
1	0.000	0.175	0.374
2	6799.627	0.638	0.751
3	8591.020	0.733	0.414
4	20,978.393	1.543	0.460
5	25,905.726	1.842	0.488
6	29,414.847	2.060	0.522
7	23,882.844	1.657	0.442
8	18,964.084	1.243	0.590
9	10,317.802	0.652	0.371
10	8845.228	0.595	0.414
11	9529.344	0.698	0.357

presents the IRF values from CO₂T in each period.

Considering the shock in TFECT, CO₂T exhibits a pronounced rise, peaking at over 20,000 by period 4, before trending downward in subsequent periods. This indicates that while increased transport energy use has a strong effect on emissions, the influence diminishes progressively as the system adjusts. Analyzing the shock in CO₂T, a self-reinforcing effect is observed. CO₂ increases at first and gradually rises to its maximum level (≈2.06 after 6 periods), after which it starts to decline. Thus, it can be concluded that the shock in CO₂ emissions causes a lasting effect, which only fades after a longer period. Regarding the shock in GREV, an initial upward response (≈0.75 in period 2) is visible, but after a few periods, the effect is ambiguous. Therefore, state budget revenues may have some impact on CO₂ emissions, but this effect is not strong.

These results suggest that CO₂ emissions can reflect changes in transport-related activity that later translates into increased energy consumption. Rather than acting as a direct driver, emissions may indicate intensified use of transport services, including freight or passenger movement. The response of GREV, although limited in magnitude, points to a indirect linkage, where emissions-related activity contributes to fiscal inflows through fuel taxation or broader economic turnover. This interpretation shows differentiated responses in timing and scale across the transport-energy-fiscal nexus.

Similarly to previous analyses, the responses to shocks in the system with GREV were examined. The impulse response from GREV is shown in Figure 18, while

Table 14. GREV’s response to shocks in the system.

Period	TFECT	CO2T	GREV
1	0.000	0.000	0.857
2	2591.472	0.274	0.407
3	7062.757	0.627	0.272
4	22,071.636	1.587	0.380
5	18,801.535	1.333	0.074
6	20,518.968	1.432	0.352
7	9682.863	0.649	0.340
8	6492.695	0.398	0.404
9	2817.795	0.192	0.186
10	8872.288	0.626	0.378
11	10,627.816	0.796	0.177

presents the IRF values from GREV in each period.

Taking into account the shock in TFECT, government revenues (GREV) rise sharply, exceeding 22,000 by period 4, and then gradually decline in the subsequent periods. This pattern highlights a direct fiscal response to increased transport energy use. Referring to the shock in CO₂T, GREV climbs more slowly, reaching approximately 1.43 in 6 period, and then experiences a downturn. This may result from cyclical changes in the domestic economy. Cyclical fluctuations are related to the degree of transport utilization, which in turn impacts carbon dioxide emissions from transport. However, these fluctuations do not cause significant disruptions in state revenue flows. When analyzing the shock in GREV, a preliminary surge in revenues (0.86 in the first period) is observed, but later oscillations occur, both upwards and downwards, which may indicate that the state’s policy in this area is not sufficiently predictable, leading to varying effects on the long-term state of funds collected in public accounts.

The results obtained by the IRF function from GREV suggest that changes in fiscal capacity, even when variable, can indirectly influence energy consumption through policy channels such as infrastructure investment or demand stimulation. These effects are not immediate but may accumulate over time, especially in sectors that are energy-intensive by nature. The reaction of CO₂T is less consistent, which aligns with the fact that emissions are not directly shaped by fiscal instruments in the polish transport sector. Overall, the findings indicate that fiscal dynamics can transmit into the energy and environmental domains, although the strength and predictability of these links depend on how public resources are allocated.

The conducted IRF analysis and observed response patterns suggest that some of the system dynamics are shaped not only by statistical interdependence, but also by underlying institutional and behavioral factors. This is particularly evident in the time-lagged and persistent reactions observed for some variables. In the case of the delayed response of government revenues (GREV) to a shock in transport energy consumption (TFECT), one explanation is the structure of the Polish tax system. While fuel excise is collected immediately upon sale, other major fiscal revenues, such as VAT or CIT, are subject to monthly or quarterly settlement cycles. As a result, it takes time for increased transport activity and fuel use to be reflected in government accounts. Additionally, rising energy consumption may stimulate broader economic activity in sectors like logistics or retail, which generates taxable value with a certain delay due to accounting practices and reporting periods. These factors help explain why the fiscal response becomes visible only after a few periods. When it comes to the prolonged effect of carbon dioxide emission shocks (CO₂T), which lasts for more than six periods, the key reasons are related to technology and cost barriers in the transport sector. Fossil fuels still dominate the energy mix in transport, and the transition to low-emission technologies, such as electric vehicles, is progressing slowly. This is due to several constraints, including: the limited availability of public charging stations (particularly outside major cities), relative high purchase costs of low-emission vehicles despite existing subsidies and also administrative delays in processing support programs. There are also behavioral aspects involved—both transport users and service providers often postpone switching to cleaner alternatives because of high up-front costs, uncertainty about future regulations, or long lifespans of current using fleets. These elements together contribute to the inertia of the system and explain why the effects of a CO₂ shock are more persistent than those of other variables. These patterns show that the dynamics captured in the model are shaped not just by statistical relationships between energy use, emissions, and fiscal outcomes, but also by real-world aspects and constraints, such as how taxes are settled or how quickly the transport sector can adapt to change.

5. Conclusions

The article analyzes the relationship between the total final energy consumption in transport (TFECT), carbon dioxide emissions from transport (CO₂T), and state budget revenues (GREV) in the context of Poland. The Johansen cointegration method and the associated test statistics, Trace and Max-Eigenvalue, were used. After identifying the existing cointegrating relationships in the analyzed time series, the Impulse Response Function (IRF) was also calculated to gain a deeper understanding of the short- and long-term dynamics between the series, as well as to identify the main transmission channels (i.e., the shocks in the system). The conducted study revealed that the total final energy consumption in transport (TFECT) is a major determinant of both carbon dioxide emissions (CO₂T) and state budget revenues (GREV). Therefore, an increase in energy consumption in transport not only leads to higher emissions but also affects the funds flowing into public accounts. The CO₂T series reacts to TFECT but also to its own shocks, which suggests a self-reinforcing effect in carbon dioxide emissions. The GREV series is strongly linked to TFECT but reacts less predictably to other time series variables. Considering the results obtained, it can be concluded that in Poland, reducing total energy consumption in transport could be an effective way to reduce carbon dioxide emissions from transport activities. Regulations and changes in fiscal policy, including fuel or emission taxes, may impact CO₂ emissions from transport, but the effect is not immediate and spreads over time. The scale of these changes, therefore, depends on long-term government strategies. State budget revenues are also significantly linked to energy consumption in transport, suggesting that tax policies related to transport activities may have a substantial impact on public finances.

The conducted analyses confirm the research hypotheses formulated in the introduction of this study. The first hypothesis, regarding the existence of long-term relationships between energy consumption, CO₂ emissions, and government revenue, is supported by the results of Johansen’s cointegration test and the Vector Error Correction Model. The second hypothesis, concerning the direction and nature of these dependencies, is also validated. The impulse response analysis (IRF) shows that an increase in energy consumption in transport leads to a rapid rise in CO₂ emissions, and affects public revenue, albeit with a certain delay. This indicates that the analyzed variables are interconnected over the long term, with varied intensities and time dynamics. These findings also have practical implications for public policy in Poland. In recent years, government initiatives have included a proposed environmental tax on internal combustion vehicles, originally scheduled to come into force in 2026, with rates depending on the EURO emission standard met by the vehicle. At the same time, increasing attention is being paid to an alternative approach emphasizing financial incentives instead of consumer costs. For instance, government subsidies of up to 40,000 PLN are currently available for the purchase of electric vehicles. Such measures are consistent with the directions outlined in the Energy Policy of Poland until 2040 (EPP2040), which emphasizes the development of electromobility, improved energy efficiency, and the decarbonization of the transport sector. An integrated fiscal and environmental approach to transport policy could play an important role in supporting these strategic goals. The results of this study can contribute to the evaluation of both cost-based and incentive-based strategies. On the one hand, increasing fuel taxes may help reduce energy consumption and CO₂ emissions while providing additional public revenue. On the other hand, incentive-based programs require careful fiscal planning but may more effectively accelerate the adoption of low-emission technologies. In light of findings, a balanced approach that combines fiscal and environmental objectives appears to be the most appropriate. Recommended directions should include: (1) fuel taxation based on emission intensity, (2) allocating transport-related tax revenues to the development of sustainable infrastructure, (3) supporting the sector’s transformation through targeted tax relief and subsidies for low-emission technologies.

In addition to general fiscal instruments, the adjustment of policy to the national institutional context is of particular importance. In the case of Poland, it is worth considering, among other things, partial redirection of fuel tax revenues to local governments. Currently, fuel tax is set centrally, but local transport is mainly financed by local governments. The introduction of a registration tax linked to CO₂ emissions could be considered (as in Germany or France), although it would be necessary to take into account the total emissions over the entire life cycle of the vehicle. It is also worth postulating better use of EU ETS funds in the transport sector, which are currently directed to a limited extent to decarbonization of transport in Poland, ensuring their greater allocation to the modernization of public rolling stock, support for electromobility or expansion of charging infrastructure. In the context of restrictions resulting from the expenditure rule and the fiscal situation, it is recommended to introduce instruments with high fiscal efficiency, i.e., those that have a low budget cost in relation to environmental effects (e.g., tax reliefs instead of direct subsidies, public-private partnerships).

This study has several limitations. First, the temporal scope of the analysis was limited to the period 1990–2022, reflecting the availability of consistent and reliable macroeconomic and sectoral data over a politically and institutionally stable period. Earlier data were excluded due to concerns about comparability and the structural differences in the economic system prior to the early 1990s. The selected time frame ensures a uniform analytical foundation based on sources from official statistics and recognized international institutions. Second, the analysis focuses on three main variables: total final energy consumption in transport, CO₂ emissions from transport, and government revenues. This choice was motivated by the need to maintain model clarity and long-term data homogeneity. Nevertheless, other factors, such as fuel prices, legal and fiscal reforms, macroeconomic shocks, or technological change, may also influence the relationships under investigation. Additionally, the impact of external shocks, such as geopolitical tensions, crises, or EU climate policies, could introduce volatility into the system that is not fully captured by the model structure. Future studies could incorporate time-varying parameters to better account for such dynamics. Third, the present analysis treats the transport sector as a whole, without disaggregating passenger and freight transport. However, these subsectors may differ in their contributions to energy consumption, emissions, and fiscal outcomes. Differentiating between them in future research could provide a more detailed understanding of the drivers and policy implications within the transport domain. Finally, while the Johansen cointegration approach used in this study is appropriate for assessing long-run relationships among integrated variables, it does not account for possible structural breaks or nonlinear effects. Given Poland’s significant institutional and policy transformations during the analyzed period, future research could benefit from applying complementary techniques, such as structural break tests (e.g., Zivot-Andrews) or ARDL-type models. These extensions would provide a broader analytical perspective while building on the solid empirical basis established in the present study. Generally, these limitations do not undermine the validity of the current findings but rather delineate the scope and applicability of the results. At the same time, they point to promising directions for future research.