Next Article in Journal
Effects of Working Capital Management on Small-Sized Businesses in Competitive Environments with Economic Policy Uncertainty—A Case Study Applied to Chilean Companies
Next Article in Special Issue
Network Integration as a Tool for Sustainable Business Development
Previous Article in Journal
Impact of Biodiversity Loss on Pollen Allergies: A Bibliometric Analysis
Previous Article in Special Issue
Simulation and Regulation of High-Quality Economic Development Based on System Dynamics and DPSIR Model
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Assessing Regional Economic Performance in Romania Through Panel ARDL and Panel Quantile Regression Models

Department of Economic Informatics and Cybernetics, Bucharest University of Economic Studies, 0105552 Bucharest, Romania
*
Author to whom correspondence should be addressed.
Sustainability 2024, 16(21), 9287; https://doi.org/10.3390/su16219287
Submission received: 23 September 2024 / Revised: 19 October 2024 / Accepted: 23 October 2024 / Published: 25 October 2024
(This article belongs to the Special Issue Sustainability in Business Development and Economic Growth)

Abstract

:
This study aims to address the persistent regional economic disparities in Romania by evaluating economic performance through Panel Autoregressive Distributed Lag (pARDL) and panel quantile regression (PQR) models. The analysis focuses on the impact of key economic variables, including research and development expenditures (CTCRD), IT infrastructures (IT), the number of universities (FCL), and the average number of employees (NMSP), on regional gross domestic product (GDPR). Using data from the Romanian National Institute of Statistics for the period 2003–2022, this research seeks to understand how targeted investments and policy interventions can stimulate growth and reduce inequalities across regions. The findings highlight the important role of R&D, IT infrastructures, and technological advancements in driving economic growth, especially in less developed areas. The study also emphasizes the importance of region-specific strategies in fostering sustainable growth, promoting economic resilience, and bridging the gap between more and less prosperous regions.

1. Introduction

Balanced regional development has been a major priority for European economic policies in terms of aiming to reduce socio-economic disparities between regions. In Romania, these disparities have significantly intensified in the post-communist period, being influenced by factors such as economic transition, inadequate infrastructures, and discrepancies in public investments. Despite the rapid economic growth of recent decades, regions in Romania continue to show significant differences in terms of economic performance and living standards. Previous research has highlighted these challenges, showing that, while economic growth has been consistent, the gap between developed and less developed regions continues to widen, primarily due to disparities in industrial development and R&D investments.
An essential aspect of regional economic development is the integration of sustainability principles into economic and development policies [1,2,3]. Investments in IT infrastructures and research and development not only stimulate short-term economic growth but also provide sustainable solutions for regions to adapt to climate change and new technological demands. A sustainable regional economy entails a balance between economic growth and environmental protection [4], contributing to the reduction in disparities and the improvement of quality of life in the long term.
Analyzing the regional disparities not only allows for a better understanding of the dynamics of the national economy but also provides a foundation for effective public policies that can stimulate economic development at the local level. In this context, advanced econometric modeling, such as pARDL and PQR, plays an essential role in assessing regional economic performance. These models allow for a detailed analysis both in the short term and in the long term, providing a clear picture of the factors that influence economic dynamics. Balan [5], in her study, highlights economic and social inequalities based on regions, emphasizing the factors that contribute to these disparities, such as economic structure and insufficient investments. Russu [6] observed, through his study, that statistical data show an increasing gap in terms of the development regions of Romania.
This study aims to address this research gap by investigating the regional economic performance of Romania, focusing on the role of R&D expenditures, IT infrastructures, education, and employment. We employ advanced econometric techniques such as pARDL and PQR to provide a nuanced analysis of both short- and long-term impacts in regard to reducing regional disparities and fostering sustainable growth. The novelty of our approach lies in its multi-factorial analysis, integrating technology, sustainable development, and innovation to offer tailored solutions for regional development.
In the context of Romania’s accession to the European Union and the efforts to reduce regional disparities [7,8], investments in infrastructures and European structural funds have played an essential role in regional development. Cohesion policies have supported the modernization of infrastructures and the economic development of less advanced regions, contributing to the reduction in disparities. This paper makes a significant contribution to understanding regional economic performance in Romania, analyzing the long-term impact of these investments and their relevance regarding European economic convergence. In the context of Romania’s EU accession and subsequent efforts to modernize infrastructures, this paper covers an essential period (2003–2022) that includes significant economic events [9] like the global financial crisis [10] and the post-crisis recovery. By assessing the impact of targeted investments on regional GDP, this research provides valuable insights into how strategic policies can promote balanced regional growth and support European convergence goals.
The structure of the paper is organized into five sections. The first section, the Introduction, establishes the regional context and the relevance of the analysis. Section two, the Literature Review, synthesizes relevant studies for the analysis of regional economics. Methodology and Data Collection, presented in the third section, describes the econometric methods used as well as the data sources. Section four, Results and Discussions, analyzes the econometric results, while the Conclusions and Policy Recommendations are presented in the final section.

2. Literature Review

The evaluation of regional economic performance has become a priority for policymakers and researchers from various economic fields. In the European Union, reducing economic disparities between regions is a central objective that is supported by cohesion and regional development strategies [11]. In this context, Romania is facing significant challenges due to pronounced economic imbalances between its regions, particularly between urban and rural areas, as well as among the historical macro-regions.
The purpose of this section is to analyze the relevant specialized literature for understanding regional economic performance, with a particular focus on the methodologies used to assess sectoral and regional impact. The selected studies provide a solid foundation for the current research, exploring various fields such as transportation infrastructure, agriculture, tourism, and energy. These studies highlight the role of development policies as well as how specific economic sectors can contribute to mitigating regional disparities.
Several recent studies analyze the aspects that contribute to the development of the regional or local economy in Romania. Bădulescu et al. [12] investigate the role of research and development in Romania’s economic growth and reveal that, although the country has experienced consistent economic growth over the past 15 years, regional disparities have deepened and investments in research and innovation remain well below the European average. Another study conducted by Kocsis [13] also emphasizes the importance of industry and continuous technological development for economic growth, demonstrating through statistical data and macroeconomic indicators that countries with a strong industrial base and focus on innovation have a competitive advantage on the international stage. His analysis highlights the economic differences between the industrialized and less industrialized regions of Romania, emphasizing that industry and research and development are the main pillars of economic growth and service development.
One of the primary goals of the EU is to reduce socio-economic disparities through various development policies. In Romania, economic disparities at the regional level are regularly assessed to identify areas in need of targeted investment. The initial aim of regional policy was to address existing regional imbalances, promote balanced development, revitalize disadvantaged areas, and encourage inter-regional cooperation. Papers by Androniceanu et al. [14] and Androniceanu and Georgescu [15] explore the concept of regional decentralization, the main types of disparities and challenges, and the specific characteristics of Romania’s macro regions. A principal component analysis is conducted to illustrate the distribution of key macroeconomic indicators and their inter-relationships. Additionally, a k-means clustering algorithm is applied to Romania’s 42 counties, leading to a classification of clusters and a comparison between clusters and living standards. This research identifies the main dilemmas and vulnerabilities of Romania’s macro regions and highlights the counties that contribute most significantly to regional imbalances. The findings underscore the need for a significant revision of development policies in Romania’s macro regions.
The study by Behrens and Thisse [16] approaches the concepts and tools of the new economic geography in order to address the existing problems in the regional economy. The authors identify two central aspects regarding what a region means and what kind of interactions we want to model and research between regions. They believe that a regional economy should be delineated from geographical, administrative, economic, or social perspectives.
In recent years, the concept of resilience has garnered increasing focus in economic geography and regional studies [17,18].
Another current study [19] focuses on the concept of regional economic resilience. The authors [19] analyze regional economies in the context of unforeseen shocks such as the COVID-19 pandemic and the energy crisis generated by the Russia–Ukraine war. The authors point out that shocks and crises can create opportunities for radical transformations in regional economies. These transformations could be more beneficial from the perspective of economic and social sustainability, providing a better trajectory than the pre-crisis one.
Gumpert [20] carries out research in which he discusses regional underdevelopment from the perspective of a Solow model that is differentiated between two regions from an economic perspective (industrial and agricultural). The study focuses on how factors such as technological progress and investment and depreciation rates influence the development of these regions, and the analysis is carried out by extending a theoretical growth model based on the Solow model. The authors suggest that the analysis of underdevelopment must include factors specific to investment and technological progress in order to better understand the divergences between regions.
Liu et al. [21] conducted a study in which they analyzed the relationship between ecology and economy using data at the administrative district level from Sichuan Province, China. Their results show that the eastern part of the province is more sustainably economically developed, while the south has better ecological quality, influenced by geographical conditions. Almost all the analyzed units show a moderate to high degree of coordination between the economy and ecology, with positive trends being seen during the period of 2010 to 2015.
Regarding innovation and local economic development, Szopik-Depczyńska et al. [22] conducted a study in which they proposed a new method for comparing regions in the EU based on their level of innovation. They used a multi-criteria method and concluded that EU regions are grouped not only based on their average development but also according to their performance in each considered area. Analyzing two case studies from a local authority in England, Ferreria et al. [23] highlight in their research how local authority managers change their roles at different stages of the analysis process for technology-based projects obtained through bidding. Ma et al. [24] demonstrate that the development of smart cities has a positive effect on regional economic growth by stimulating urban innovation capacity and entrepreneurship, being essential in regions with a high market degree. Additionally, Shi et al. [25] emphasize the importance of innovation for sustainable growth, highlighting spatial externalities and their impact on domestic demand. On the other hand, Pusz et al. [26] analyze the circular economy from a social and relational perspective, highlighting the role of social enterprises in local socio-economic integration, particularly through the activities of reuse and upcycling. These studies provide a solid foundation for understanding the interaction between innovation, sustainable development, and the circular economy. Milin et al. [27] investigate the impact of foreign direct investment, trade, final consumption, exports, and imports on economic growth in Romania, using annual data from the period 1990–2020. The analysis highlights that final consumption and trade have a positive impact on economic growth in both the short and long term, while exports of goods and services significantly contribute to economic growth through positive shocks. However, foreign direct investments and imports have had a negative effect on economic growth in certain contexts, indicating the need for prudent policies to mitigate uncertainties related to trade and foreign investments. The analysis of authors could also have implications for the regional and local economy of Romania, as trade and foreign direct investments are essential for the economic dynamics at the regional level.
The study conducted by Chriță et al. [28] provides a valuable perspective on the development of academic publications in economic cybernetics, with relevant implications for regional and local economies. The results highlight essential themes such as community-level economic growth and local resilience, key factors in regional economic development. In particular, the analysis highlights policies tailored to regional specifics, emphasizing the need for customized approaches to support the resilience and adaptability of local economies to global changes, such as the COVID-19 pandemic. This connection between economic cybernetics studies and their impact on the local economy demonstrates how performance optimization and risk management at the regional level can support sustainable development and competitiveness in the context of volatile economic circumstances.
The study conducted by Gurgu et al. [29] analyzes how creative sectors, such as visual arts-, design-, and technology-based industries, contribute to the sustainable development of Romania. This shows that creative industries stimulate innovation, economic growth, and social cohesion. The research emphasizes the importance of integrating cultural policies with innovation strategies to promote sustainable development. Creative sectors can become catalysts for regional and local development, contributing to the creation of entrepreneurial ecosystems and technological progress.
Table 1 presents an evaluation of studies specific to the local and regional economy. We can see that the identified studies cover various economic fields: transport (ports, railways), agriculture, tourism, energy, and water policies. This provides a broad perspective on sectoral economic impacts at the regional level. Also, the studies focus on several countries and regions, offering an interesting comparison between regional economies at the global level. Most studies use advanced econometric methods (CGE, ARDL, SEM), which gives them rigor and reliability in the assessment of regional and local economic impacts. These methodologies allow for both short-term and long-term analysis, providing a complete view of economic effects. The articles in Table 1 were selected based on their high number of citations.

3. Methodology and Data Collection

This study employs two complementary econometric techniques: the pARDL model and PQR. The pARDL approach is used to examine both short-run and long-run dynamics among the economic variables (CTCRD, IT, FCL, NMSP, and GDPR) across Romania’s regions over the period 2003 to 2022. The ARDL method allows for the identification of long-term equilibrium relationships while capturing short-term fluctuations, being useful when variables exhibit different levels of integration. To complement the dynamic insights from pARDL, PQR is applied to investigate how the impact of these variables varies across different levels of economic development. By dividing regions into quartiles based on their economic performance, this technique allows for a more detailed analysis of distributional effects, highlighting how the influence of factors like CTCRD, IT, and workforce expansion differ between more developed and less developed regions. Quartile regression offers a deeper understanding of the heterogeneous effects across the economic spectrum, which helps in identifying disparities and regional differences in economic performance.
The pARDL model provides a temporal dimension by identifying the dynamic interactions between variables over time, revealing both short-term adjustments and long-term equilibrium relationships. On the other hand, PQR adds a distributional perspective by showing how these relationships vary across different economic strata, particularly among regions with differing levels of development. Together, these techniques offer a comprehensive evaluation of Romania’s regional economic performance by combining insights into both time-dependent effects (through pARDL) and distributional impacts (through quartile regression). This dual approach ensures a nuanced understanding of how regional policies should be tailored to address both long-term growth dynamics and regional disparities.
Figure 1 illustrates the main stages of the methodology applied in the research. The process begins with defining the narrative objective of the scientific investigation, followed by data collection from the National Institute of Statistics in Romania. Preliminary statistical analysis includes descriptive methods, Pearson, Spearman, and Kendall correlations, as well as the covariance matrix. Subsequently, the stationarity of the data is tested through panel unit root tests, and the dynamic relationships between variables are assessed using the pARDL model. The pARDL model is particularly suitable for this analysis because it allows for examining both short- and long-term relationships while handling variables integrated at different orders (I(0) and I(1)). This makes it versatile for mixed datasets typically encountered in regional economic studies. Furthermore, the model includes an Error Correction Term (ECT) which captures the speed at which deviations from the long-run equilibrium are corrected, providing insights into the resilience and adjustment process of regional economies. On the other hand, PQR is used to provide a more comprehensive view of the data, allowing us to analyze how the effects of independent variables vary across different quantiles of the dependent variable. This approach captures heterogeneous impacts that may not be evident through traditional mean-based methods, making it valuable for understanding the distributional effects of R&D expenditures, IT infrastructures, and other factors. Also, PQR provides a deep understanding of the distribution of dependent variables, while the Granger causality test examines the direction of the relationships between variables. The research concludes with findings and policy recommendations based on the results of the analysis.
Each step in the methodological process contributes to the study’s objective of assessing the factors influencing regional economic performance. The preliminary statistical analysis ensures that the data meet the requirements for robust econometric modeling, while unit root testing establishes the stationarity properties needed for the pARDL model. The use of PQR enables a nuanced exploration of how different factors impact economic outcomes across various levels of regional development, offering more tailored policy insights.

3.1. Data Stationarity

The data stationarity is tested by means of the following panel unit root tests: the Levin, Lin and Chu (LLC) test [40], the Im, Pesaran and Shin (IPS) test [41] and the ADF-Fisher Chi-square (ADF) test [42]. Payne et al. [43] also conduct a study in which they use a panel unit root test to examine the stationarity of electricity consumption in India. These tests generally offer greater power than unit root tests applied to individual time series, as they enhance the number of observations by pooling time series data from multiple units. Next, we will overview these three panel unit root tests. For a clearer understanding of all the acronyms used in our study, see Abbreviations, which describes these acronyms.

3.1.1. LLC Panel Unit Root Test

The LLC test enhances the power of unit root testing by combining individual time series rather than testing each one separately. The LLC panel unit root test was also used by Babar et al. [44] to assess the relationships between macroeconomic forces and the stock markets in Asia for the period 1999–2019.
It accounts for cross-sectional dependencies and time dynamics, making it particularly useful for datasets where cross-sectional units are expected to exhibit similar stochastic trends. Levin et al. [40] developed this panel unit root test, which is based on an augmented Dickey–Fuller (ADF) model, as outlined in the following basic model:
X i t = α i + β i X i t 1 + δ i t + j = 1 k γ i j X i t j + ε t
In Equation (1), is the difference operator, X i t is the dependent variable i at time t , and ε t is the error term. The null hypothesis is H 0 :   β 1 = β 2 = = β = 0 , versus the alternative hypotesis H 1 :   β 1 = β 2 = = β < 0 .
This approach has also been used by other authors in their studies [45,46,47].

3.1.2. IPS Panel Unit Root Test

The IPS test allows for heterogeneity in the autoregressive coefficients across different cross-sections, unlike the LLC test, which assumes a common autoregressive parameter across all units. The IPS test averages the individual Dickey–Fuller (or augmented Dickey–Fuller) test statistics from each series in the panel (Equation (1)). This makes the IPS test more adaptable when the LLC test’s assumption of homogeneity is too restrictive. The flexibility of the IPS test is particularly advantageous in panels where individual units might experience different shocks or dynamics that need to be considered. The IPS test, like standard unit root tests, tests the null hypothesis of a unit root but with the alternative hypothesis that at least one cross-sectional unit in the panel is stationary, differing from the LLC’s assumption of collective stationarity.
The null hypothesis of the IPS test is H 0 :   β 1 = β 2 = = β = 0 versus H 1 :   some but not necessarily all β i < 0 . The IPS test computes the average of the individual t-statistics. The t-bar statistic is then used to assess the order of integration under the null hypothesis ( β i = 0 ):
t ¯ = i = 1 N t β i N
The t-statistic is obtained from the panel cross-sections, and N is the sample size. t ¯ follows a normal distribution. Consequently, t ¯ will be converted into a standardized normal z-bar statistic, z ¯ :
z ¯ = N t ¯ E t ¯ β i = 0 V a r t ¯ β i = 0
With E t ¯ β i = 0 and V a r t ¯ β i = 0 the mean and variance of t β i , obtained by Monte Carlo simulations [42].

3.1.3. Fisher-Type Panel Unit Root Tests

The Fisher-type panel unit root tests are employed to assess the presence of a unit root in panel data. Originally developed by Dickey and Fuller [48], and later adapted by Maddala and Wu [42] and Choi [49] for panel data, these tests combine the p-values from individual unit root tests across different cross-sectional units (such as countries) into a single test statistic. This approach improves the reliability of inferences from panel data by accommodating varying dynamics and incorporating information from multiple sources. It is particularly valuable because it allows for heterogeneity among individual series within the panel, recognizing that different units may have distinct stochastic properties.

3.2. Panel ARDL

The panel Autoregressive Distributed Lag (pARDL) model is used to examine the dynamic relationship between variables over time across different entities [50,51,52]. It is especially well suited for panel data, which consist of observations spanning multiple time periods for various cross-sectional units. An advantage of the pARDL model is its ability to handle variables integrated in different orders (I(0), I(1), etc.) [53], allowing it to accommodate both stationary and non-stationary data. Additionally, the model includes an Error Correction Term (ECT), derived from the long-term relationship, which measures the speed at which variables return to equilibrium following a disturbance.
By Pesaran and Shin [54], the ARDL (p, q) model is specified as:
Y i t = j = 1 p φ i j Y i t j + j = 0 q δ i j X i t j + ϑ i + ϵ i t
i = 1 , . . . , N is the cross-section index, t = 1 , . . . , T is the time index, X i t represents the independent variables, ε i t is the error term, ϑ i is the fixed effect, and p and q are lag numbers. The Error Correction Model is:
Y i t = Φ i Y i t 1 + θ i X i t + j = 1 p 1 φ i j Δ Y i t j + j = 0 q 1 δ i j Δ X i t j + ν i + ϵ i t
Φ i is ECT and θ i depicts the long-run equilibrium between Y i t and X i t . φ i j and δ i j are short-run coefficients. Y i t and X i t are cointegrated if the ECT is between −1 and 0 and statistically significant.
The Error Correction Term (ECT) is an important component that measures the speed of adjustment back to the long-run equilibrium following short-term disturbances. A statistically significant ECT between −1 and 0 indicates that, after a shock, regional economies gradually return to equilibrium. This feature of the ARDL model allows us to assess the resilience of regions in adjusting to changes in R&D investments or policy interventions, providing insights into the long-term stability of economic growth.
Recent studies have also employed the pARDL model to investigate dynamic relationships in regional economic contexts, demonstrating its robustness in capturing both short- and long-term effects across diverse datasets [37,55,56,57,58].

3.3. Panel Quartile Regression

PQR extends the traditional econometric analysis by allowing for the estimation of effects across different quantiles of the dependent variable’s distribution. This approach is particularly relevant for the current study, as regional economic performance can be heterogeneous, with the impact of predictors like R&D spending or IT infrastructures varying depending on the level of regional development. By focusing on multiple quantiles (e.g., median, 25th percentile, 75th percentile), we obtain a more detailed understanding of how these factors influence economic outcomes, which is essential for formulating differentiated policy recommendations.
Quantile regression, first introduced by Koenker and Bassett [59] in 1978, extends the classical least squares approach, which focuses on estimating conditional mean models, to allow for the estimation of multiple models corresponding to different conditional quantile functions. A key example is the median regression estimator, which is obtained by minimizing the sum of absolute errors.
PQR is a statistical method used to analyze the relationship between variables in panel data, with a focus on understanding how the effects of independent variables vary across different points (quantiles) of the dependent variable’s distribution [60,61]. This approach is particularly useful when the impact of predictors is not uniform across all levels of the outcome variable, allowing for a more detailed and nuanced understanding of the data. Traditional methods like ordinary least squares (OLS) regression estimate the average effect of independent variables on the dependent variable. However, this average effect may not capture the variation in effects across different levels of the dependent variable. PQR, on the other hand, allows for the estimation of effects at various quantiles (e.g., median, 25th percentile, 75th percentile) of the dependent variable’s distribution [62]. This means that you can see how the relationship between the independent and dependent variables differs at different points in the outcome distribution, which is important in understanding heterogeneous impacts. Quantile regression is more robust to outliers than mean regression techniques. By focusing on different quantiles, it is less influenced by extreme values, making it a valuable tool when the data contain outliers or when the distribution of the dependent variable is skewed. This method is particularly useful in understanding complex economic relationships where the impact of predictors may not be uniform across all levels of the outcome, offering deeper insights and more tailored policy recommendations.

3.4. Granger Panel Causality Test

The Granger panel causality test [63] is an extension of the Granger causality test applied within the context of panel data, which consists of observations across multiple time periods and cross-sectional units. This test is used to determine whether one time series can predict another, indicating a potential causal relationship between the variables. In a panel setting, the Granger panel causality test evaluates the direction of causality across the entire panel, taking into account both time dynamics and cross-sectional dependencies. It can test whether changes in one variable consistently precede changes in another variable across the units in the panel, suggesting causality.
The combination of pARDL, PQR, and Granger panel causality test methods provides a comprehensive framework for analyzing regional economic performance. The ARDL model captures both short- and long-term dynamics, including the adjustment towards equilibrium, while quantile regression offers insights into how the impact of different variables varies across regions with different levels of economic development. Additionally, the Granger panel causality test is used to identify the direction of causality between variables, allowing us to understand not only the strength and nature of relationships but also the potential causal pathways. This multi-method approach enables a nuanced analysis of how factors such as R&D expenditures, IT infrastructures, and labor force size drive economic growth while also addressing the distributional effects and temporal dynamics across regions.

4. Results and Discussions

4.1. Input Data Analysis

In Table 2, the logarithmic variables used for regional economic data come from the statistical sources of the Romanian National Institute of Statistics [64], reflecting relevant economic aspects. These include variables like regional gross domestic product (GDPR), research and development expenditure (CTCRD), IT equipment in the administration of public and private university education units (IT), the number of universities (FCL), and the average number of employees (NMSP). Most of the variables show some deviations from normality, particularly CTCRD and NMPS, which have significant skewness and kurtosis. IT also shows a significant deviation, but this is primarily due to its platykurtic nature. CTCRD and IT show relatively higher variability (standard deviation), suggesting more dispersion in terms of R&D expenses and IT equipment across regions. NMPS, with the lowest standard deviation, shows the least variability. CTCRD and NMPS have a notable right-skewness, indicating that a few high-value observations are pulling the mean to the right. IT’s platykurtic distribution suggests a flatter spread with less extreme values compared to a normal distribution. In economic terms, these distributions indicate that, while regional GDP, number of universities, and average employment are relatively stable, there is more variability in R&D spending and IT infrastructures, with potential outliers influencing the overall trends in these variables. This variability could reflect differences in regional priorities, funding availability, or economic development strategies.
In Figure 2, a multidimensional analysis of the correlations between the analyzed indicators and their covariance matrix was conducted to understand the relationships and associations between these regional economic indicators. For example, we observe a Pearson correlation value of 0.49 between GDPR and CTCRD, indicating a moderate positive correlation between regional GDP and expenditures in research and development activities. This suggests that, as research and development expenditures increase, regional GDP tends to grow to a moderate extent. Additionally, the Spearman (0.48) and Kendall (0.39) correlations between these two indicators confirm the existence of a similar relationship, highlighting that the relationship between GDPR and CTCRD is consistent though not necessarily linear. The covariance of 0.32 indicates that the two variables move together, although the magnitude is not very high.
In Figure 3, we observe the evolution of regional GDP across various regions of Romania from 2003 to 2022. There is a clear trend in consistent GDP growth across all analyzed regions over nearly two decades. The Bucharest-Ilfov region stands out with a consistently higher regional GDP compared to other regions, suggesting that it serves as the economic engine of Romania. Regions like Center, West, and South-Muntenia also show relatively high GDP levels compared to others, which may indicate an unequal distribution of economic development across the country. In some regions, such as South-Muntenia and South-West Oltenia, there are periods of stagnation or even slight declines in GDP in certain years (e.g., 2015 in South-Muntenia or 2013 in South-West Oltenia). These fluctuations could be linked to various local, national, or global economic events, such as the 2008–2009 financial crisis. Overall, there is a tendency towards convergence, where less developed regions (like the North-East) are beginning to close the gap with more developed regions (like the Bucharest-Ilfov region). Although disparities remain, the growth rate in less developed regions is encouraging.
Across most regions, there is a noticeable upward trend in both GDPR, and, according to Figure 4, there has been a noticeable trend in increasing research and development expenditures across most regions over the last two decades. This growth reflects an increased focus on innovation and technological development, which are essential factors for long-term economic competitiveness. The Bucharest-Ilfov region stands out with the highest R&D expenditures throughout the analyzed period. Similarly, the West and Center regions also show substantial investments in R&D, indicating strong potential for innovation and economic growth. In contrast, the North-West and North-East regions exhibit more modest increases in R&D spending compared to Bucharest-Ilfov and other more developed regions. However, these regions continue to invest in research, which could lead to future economic developments.
The data also highlight periods of economic resilience, particularly during the global financial crisis of 2008–2009. For instance, in the Bucharest-Ilfov region, there were slight decreases in R&D expenditures during those years, but the region managed to maintain a high level of investment in this sector, underscoring its economic robustness. Additionally, the positive correlation between GDPR and CTCRD observed in Figure 2 suggests that regions with higher R&D expenditures tend to have higher regional GDP, as seen in the Bucharest-Ilfov, West, and Center regions. This indicates that investments in research and innovation are important drivers of economic growth. Therefore, while Bucharest-Ilfov stands out as a major center of innovation, other regions demonstrate growth potential, albeit at different rates. Continued investments in research and development will be essential for sustaining this growth momentum and reducing regional disparities.
Although there has been a gradual increase in R&D spending across most regions, many still have not returned to the levels observed in 2003. This may be due to persistent disparities in the allocation of R&D resources, which tend to favor more developed regions such as Bucharest-Ilfov. Regions like the North-West and North-East face structural challenges, including a lack of research infrastructure, lower access to skilled labor, and economic constraints that limit the capacity to absorb and utilize R&D investments effectively. Consequently, while positive growth trends are observed, closing the gap with more developed regions remains a challenge that requires targeted policy interventions to ensure balanced regional development.
The significant drop in R&D expenditures in 2005, as observed in Figure 4, can be explained in the context of Romania’s accession to the European Union. During that period, Romania underwent a process of aligning with EU requirements which involved structural reforms and changes in budget allocation. For instance, the 2003 Regular Report on Romania’s progress towards accession [65] highlighted the need for economic adjustments to meet integration criteria. This has likely led to the reallocation of funds to other priority areas, such as infrastructures or institutional reforms, which temporarily affected investments in R&D. Moreover, economic restructuring and privatization efforts during that time influenced the economic environment, causing fluctuations in R&D financing. Nevertheless, the gradual increase in subsequent years indicates a recovery and recalibration of R&D investments, as Romania has moved closer to completing the accession process and has benefited from new EU development funds.
In Figure 5, the evolution of IT equipment in the administration of public and private university education units across different regions of Romania from 2003 to 2022 is presented. In most regions, a consistent increase in IT equipment can be observed over the two decades. This trend reflects a growing emphasis on integrating technology into higher education, which is essential for modernizing the educational process and aligning it with the demands of the 21st-century labor market. The Bucharest-Ilfov region stands out with the highest levels of IT equipment throughout the analyzed period. This underscores the importance of the capital as a center of innovation and technology where IT infrastructure is essential for supporting high-quality education. In the North-West region, a steady increase in IT equipment is also observed, indicating continuous adaptation to new technologies in education as this region develops its technological infrastructure. Regions such as South-Muntenia and South-West Oltenia show lower levels of IT equipment, but a gradual increase is evident. Overall, the data indicate continuous technological adaptation, even during periods of economic crisis. For example, during the global financial crisis of 2008–2009, most regions continued to increase or maintain their levels of IT equipment, highlighting the importance of technology in maintaining the quality of education.
From the perspective of the evolution of the number of public and private universities at the regional level in Romania, Figure 6 shows that, in the North-West region, the number of universities remained relatively constant over the period 2003–2022. After a period of decline between 2007 and 2010, the FCL indicator stabilized at slightly lower values than the initial ones. The Center region experienced a gradual decrease in the number of universities after 2005. Although there were a few slight increases, they were not sufficient enough to reverse the overall downward trend. The South-Muntenia region presents notable variations, with a significant decrease in 2007, like other regions. The Bucharest-Ilfov region, being a major university center, maintained the highest level of the indicator, although a steady decline has been observed since 2003.

4.2. Empirical Results

This section presents the empirical findings of the study, focusing on the relationships between key economic variables and regional economic performance in Romania. Using the pARDL and PQR models, we aim to assess both the short-term and long-term effects of CTCRD, IT, FCL, and NMPS on regional GDP. The analysis begins with the results of the unit root tests to determine the stationarity of the data, followed by an evaluation of the dynamic relationships through the pARDL model. The subsequent quantile regression analysis provides insights into how these relationships vary across regions with different levels of economic development. This approach allows us to capture the heterogeneous effects of the variables, offering a comprehensive understanding of the factors driving regional economic performance.
The linear dependence in our model is:
G D P R = f ( C T C R D , F C L ,   I T ,   N M P S )
From now on we will work with the natural logarithm of the variables.
G D P R i t = + β 1 C T R D i t + β 2 F C L i t + β 3 I T i t + β 4 N M P S i t + ε i t
The data stationarity is tested by applying the following panel unit root tests: LLC, IPS, and ADF, as seen in Table 3.
Given that some variables, such as GDPR and CTCRD, are non-stationary at levels but become stationary after first differencing, the pARDL model is well suited for our analysis because it can handle variables integrated in different orders (I(0) and I(1)). The ARDL approach allows us to model both the short-run dynamics and long-run relationships simultaneously. However, to address potential concerns about cointegration, we have also included an Error Correction Term (ECT) in our model which captures the adjustment speed towards a long-run equilibrium. To strengthen the robustness of the analysis, we considered alternative approaches, such as the Vector Error Correction Model (VECM), and found consistent results, reinforcing the validity of using ARDL for this study.
From Table 4, it follows that all the variables become stationary after differencing once. Next, we apply pARDL for the economic macro-region.
The long-run dynamics in the pARDL model show that, in the long term, increases in CTCRD spending have a small positive effect on GDPR. However, this effect is not statistically significant, implying that R&D expenditures might not have a strong or consistent impact on regional economic growth across Romanian macro regions. This could indicate inefficiencies in how R&D investments are being utilized or the time lag required to see substantial benefits from such investments.
The negative coefficient (−1.09) implies that an increase in the FCL is associated with a decrease in GDPR in the long run. Considering the negative and significant coefficient of the FCL variable, the reduction in economic growth can be explained from multiple perspectives. First, the effect of this variable may depend on its interaction with other economic variables or the specific context of the region. For example, in economies oriented towards industries that require skills developed in universities, the impact of FCL might be more pronounced. An increase in the number of universities may lead to an oversupply of graduates, creating a risk of underemployment. The quality of education may vary between newly established and more established institutions, which could affect graduates’ readiness for the labor market. Additionally, expanding the number of universities could dilute financial and intellectual resources, impacting the development of programs and research initiatives. In less developed regions, the impact of universities might be limited if complementary factors like infrastructure or research and development investments are lacking. In conclusion, the negative effect of the FCL variable may reflect a misalignment between educational outcomes and regional market demands or an inefficient allocation of educational resources.
The positive and significant coefficient (0.53) indicates that investment in IT infrastructures within universities has a strong positive impact on long-term economic growth. This suggests that modernizing educational facilities with advanced IT equipment enhances productivity and innovation, thereby contributing to regional economic development.
The coefficient of 4.40, which is highly significant, indicates that a larger workforce is strongly associated with a higher GDPR. This underscores the importance of labor as a key driver of economic growth, with more employees likely contributing to increased productivity, consumption, and overall economic activity in the regions.
The Error Correction Term is negative and significant (−0.32), indicating that, if GDPR deviates from its long-run equilibrium, about 32% of this deviation is corrected in the next period. This shows that the regional economies have a relatively strong mechanism to revert to their long-run growth path after short-term shocks.
The negative coefficient (−0.36) of D(GDPR(−1)) suggests that high growth in the previous period could lead to slower growth in the current period, possibly due to diminishing returns or cyclical adjustments in the economy.
The short-run coefficients for R&D spending are negative, which is surprising and suggests that increasing R&D spending may initially slow economic growth. This could be due to the time lag between R&D investment and the realization of its economic benefits or inefficiencies in the allocation of R&D funds. The pARDL model used in this study is particularly well suited for distinguishing between short-run and long-run effects. The negative short-run impact of R&D spending reflects the fact that investments in research and development typically take time to yield positive economic outcomes. This is consistent with the nature of R&D, where immediate returns are not expected due to the time required for innovations to materialize, technologies to be adopted, and skills to adjust. The ARDL model captures these delays through its long-run equation, which models the equilibrium relationship between R&D spending and regional economic performance over time. In our results, the long-run effect of R&D spending is positive, albeit statistically insignificant, suggesting that, once these investments fully translate into economic activities, they contribute positively to growth. The Error Correction Term (ECT), which is highly significant and negative (−0.32), indicates that deviations from the long-run equilibrium are corrected over time, supporting the idea that the economy adjusts to R&D investments even though the benefits may not be immediate. Therefore, the ARDL approach fundamentally includes a dynamic error correction mechanism that adjusts short-run imbalances toward the long-run equilibrium, effectively addressing the reviewer’s suggestion to apply dynamic error corrections. The model’s ability to differentiate between the immediate adjustments to changes in R&D spending and the accumulated long-run impact is one of its main strengths, allowing for a more comprehensive understanding of the relationship between R&D and economic growth. The short-run impact of IT investment is negative, which might indicate that initial investments in IT infrastructures disrupt existing processes or require time for integration and effective use. However, this negative impact diminishes over time, as indicated by the less significant lagged effect. An increase in employment has a positive short-run impact on GDPR, reflecting immediate gains from higher productivity and spending. However, the lagged negative effect, though not significant, suggests that there could be short-term adjustments or inefficiencies that temper this growth.
The short-run effects of increasing the FCL are negative, especially with a significant lagged effect. This may reflect the immediate costs associated with expanding educational institutions (e.g., construction, staffing) without immediate economic returns. The results obtained from Table 5 provide valuable information about the impact of variables on economic performance at this level.
CTCRD expenses have a positive effect on the lower quartile of GDPR, but this effect is not strong or significant. This suggests that, in regions with a lower GDPR (25th percentile), CTCRD expenses do not have a significant impact on economic output. The number of universities has a strong and significant negative impact on GDPR at the lower end of the distribution. This might indicate that, in less economically developed regions, an increase in the number of universities could be associated with inefficiencies or high costs relative to economic output.
IT equipment has a significant and strong positive effect on GDPR at the 25th percentile, suggesting that technological investments are important for boosting economic performance in less prosperous regions. The number of employees also has a significant positive impact on GDPR, indicating that workforce size is an important driver of economic output in these regions. The model explains approximately 53.7% of the variance in GDPR at the 25th percentile, indicating a relatively good fit for this quantile.
In Table 6, the results of the quantile regression for the median quartile (50%) are presented.
CTCRD expenses now have a significant positive impact on GDPR at the median level, suggesting that, as regions become more economically developed, CTCRD investments start to play a more essential role in driving economic growth. The negative effect of the number of universities remains significant but is less pronounced than at the 25th percentile. This could suggest that the initial negative impact of expanding university infrastructure diminishes as the region’s economy grows. IT equipment continues to have a strong positive effect on GDPR at the median, although the impact is slightly less than at the 25th percentile. The positive effect of the number of employees on GDPR remains strong, though slightly reduced compared to the 25th percentile. The model explains about 50.2% of the variance in GDPR at the median, showing a slightly lower fit compared to the 25th percentile.
In Table 7, the results of the quantile regression for the 75th quartile are presented.
CTCRD expenses have the strongest positive and significant impact on GDPR at the 75th percentile. This indicates that, in more economically advanced regions, R&D is a critical driver of further economic growth. The negative effect of the FCL continues to decrease, indicating that the adverse effects associated with university expansion are even less pronounced in wealthier regions. The positive impact of IT equipment on GDPR, while still significant, is weaker at the 75th percentile compared to lower quantiles, possibly due to diminishing returns in more advanced regions. The positive effect of the number of employees on GDPR remains significant but is the lowest among the quantiles, suggesting that, as regions become more developed, the marginal impact of additional employees decreases. The model explains about 48.7% of the variance in GDPR at the 75th percentile, showing the lowest fit among the three quantiles. The relatively low adjusted R-squared values in our PQR models are typical for such analyses, as quantile regression does not aim to maximize explained variance but rather to estimate the conditional distribution of the dependent variable. This approach provides valuable insights into how the effects of predictors differ across various levels of economic performance.
The bias-corrected linear dynamic panel data estimation technique was applied as a robustness check in the next stage to forecast the long-run effects of the model’s variables. Table 8 presents the results of a bias-corrected linear dynamic panel data model, examining the long-run impact of various factors on GDPR. CTCRD has a positive but small effect on GDPR, and the effect is statistically significant at the 1% level. This suggests that increased spending on research and development activities is positively associated with GDPR in the long run, though the magnitude of this effect is relatively modest. FCL has a negative and statistically significant impact on GDPR at the 1% level. This suggests that an increase in the number of universities is associated with a decrease in GDPR, which may indicate issues such as inefficiencies or resource allocation challenges in regions with a higher number of universities. IT equipment in educational administration has a negative but statistically insignificant effect on GDPR. This implies that IT investments in this context may not have a meaningful impact on GDPR or that the effect could be neutral. NMPS has a strong and statistically significant positive effect on GDPR at the 1% level. This indicates that regions with a higher average number of employees experience a significant increase in GDPR, highlighting the importance of labor force size in driving regional economic performance.

4.3. Comparison across Quantiles

The positive impact of R&D expenses on GDPR increases as you move from the 25th to the 75th percentile. This implies that R&D is more effective in stimulating economic growth in regions with a higher GDPR. The negative impact of the number of universities on GDPR weakens as you move up the distribution. In less developed regions (25th percentile), expanding university infrastructures might be associated with inefficiencies or higher costs relative to the economic benefits. However, as regions become more developed, these negative impacts diminish. IT equipment consistently boosts GDPR across all quantiles, though its impact decreases slightly in more prosperous regions (75th percentile). This suggests that, while IT investments are essential for economic growth, the relative benefit may be lower in regions where GDPR is already high. The number of employees has a strong positive effect on GDPR across all quantiles. However, its impact slightly decreases as regions become more economically advanced, indicating potential diminishing returns to labor in wealthier regions.
The next step involves conducting a causality test (Table 9), as outlined by Granger [63], to identify the direction of causality.
Increased CTCRD spending leads to higher regional GDP. Economically, this suggests that investments in research and development activities spur economic growth. R&D can lead to innovations, technological advancements, and increased productivity, all of which contribute to higher economic output. This aligns with economic theories that emphasize the importance of innovation and technology in driving long-term economic growth. A higher regional GDP leads to increased CTCRD spending. This reverse causality indicates that, as regions become wealthier, they have more resources to allocate toward R&D. Economic growth generates more revenue, enabling both public and private sectors to invest more in research and innovation, creating a virtuous cycle of growth and innovation.
An increase in the number of universities leads to a higher regional GDP. Economically, this suggests that universities contribute to economic growth by educating the workforce, fostering innovation, and attracting research funding. Universities can also drive regional development through knowledge transfer, entrepreneurship, and creating a skilled labor force that enhances productivity. Higher regional GDP leads to an increase in the number of universities. As regions grow economically, there is likely more demand for higher education and the financial means to establish or expand educational institutions. A higher GDP also increases government and private investment in education infrastructures, further reinforcing economic growth.
A higher regional GDP leads to increased investment in IT equipment for universities. Economically, this indicates that wealthier regions are better able to invest in modern educational infrastructures, including technology. This investment can enhance the quality of education, making universities more effective in producing skilled graduates and in conducting cutting-edge research, which in turn supports economic growth.
An increase in employment leads to a higher regional GDP. This is a classic economic relationship where a larger workforce contributes directly to higher production and economic output. More employees generally means that more goods and services are produced, leading to economic growth. A higher regional GDP leads to increased employment. As the economy grows, businesses expand, leading to more job creations. Economic growth often generates demand for labor, which further supports the economy through increased consumer spending and investment.
The interaction between the number of universities and CTCRD spending suggests that educational institutions are key players in driving research and development. More universities can lead to more R&D activity, and vice versa.
A relationship where IT investments drive CTCRD spending suggests that modern technological infrastructures are essential for effective research and innovation. However, the reverse is not significant, implying that CTCRD does not necessarily lead to immediate IT upgrades.
The finding that employment levels Granger-cause CTCRD expenses underscore the role of a skilled workforce in driving research activities. More workers, particularly in knowledge-intensive industries, are likely to contribute to higher levels of innovation and CTCRD.

5. Conclusions and Policy Recommendations

The results of this study offer important insights into the factors driving regional economic performance in Romania.
The pARDL model indicates that, while research and development expenditures have a positive impact on long-term growth, the short-term benefits may be limited if investments are not efficiently managed or aligned with market demands. This suggests that strategic planning is needed to ensure that R&D projects are relevant to regional economic priorities, with a focus on sectors that can rapidly translate research outcomes into practical applications. The negative impact of the number of universities on GDP highlights potential mismatches between educational expansion and labor market needs, suggesting that expanding higher education must be accompanied by improvements in quality and relevance to local economic conditions.
Investments in IT infrastructure and a larger workforce emerge as key drivers of economic growth, with IT having a significant impact across all regions. This underscores the importance of technological modernization in fostering productivity and innovation. However, the initial disruptions associated with IT investments should be addressed through adequate support measures, such as training programs for staff to adapt to new technologies. Policymakers should prioritize quality improvements in education and the alignment of academic programs with market needs, as well as encourage targeted investments in IT to maximize economic benefits.
The quantile regression analysis further reveals that the impact of CTCRD becomes increasingly important as regions develop, indicating that R&D investments are more effective in stimulating economic growth in more prosperous regions. The decreasing negative impact of FCL suggests that the challenges associated with university expansion are particularly significant in less developed areas, where complementary infrastructure and resources may be lacking. Consequently, policies aimed at higher education expansion should be tailored to regional conditions, with an emphasis on quality and resource allocation. The role of IT as a growth driver is evident across all regions, though its marginal benefits may decrease in the most advanced areas, suggesting that continued investments in digitalization are needed to maintain growth momentum.
The results of the Granger panel causality test suggest a dynamic and interdependent relationship between human capital (universities and employees), technological infrastructure (IT), research and development, and economic growth. Regions that effectively invest in these areas are likely to experience a self-reinforcing cycle of growth, where improvements in one area stimulate advancements in others, collectively driving the economy forward. This highlights the need for a holistic approach to economic development where policies in education, technology, and innovation are closely coordinated to achieve sustainable growth.
The results obtained in this study generate a few policies that we can recommend. Given the positive long-term impact of CTCRD on regional GDP, especially in more developed regions, the government should prioritize R&D funding in strategic sectors that align with regional economic strengths. In less developed regions, targeted incentives could stimulate private-sector investment in innovation and create an environment conducive to technology transfer. Investments in IT infrastructure should be a priority in terms of accelerating economic growth, particularly in sectors such as education, healthcare, and public administration. Digitalization initiatives should include not only hardware investments but also training programs to enhance digital skills across the workforce. Educational policies should focus on aligning university curricula with the needs of the local economy. This may involve creating partnerships between universities and industries to develop programs that address skill gaps and promote entrepreneurship. Policymakers should implement support measures, such as vocational training and reskilling programs, to mitigate the short-term disruptions associated with the adoption of new technologies. This will help ensure that IT investments translate into productivity gains more quickly.
As in any scientific study, it is important to acknowledge the potential limitations of the work. The data used are at a macro-regional level, which may obscure intra-regional disparities and localized economic dynamics. Additionally, the ARDL model does not account for indirect effects, such as educational policies influencing economic outcomes through channels like labor market dynamics or migration patterns. Future research should consider using more granular data to capture within-region variations and explore alternative econometric techniques to address endogeneity issues more effectively. While multicollinearity and endogeneity can pose challenges in terms of econometric analysis, we took steps to minimize these potential issues. The choice of variables was carefully considered based on theoretical foundations and prior empirical research, ensuring that they capture distinct aspects of regional economic performance. Additionally, we applied robustness checks, including a bias-corrected linear dynamic panel data estimation, to validate the stability of our results. Future research could explore using instrumental variables or other advanced econometric techniques to further address endogeneity concerns.
Also, future research directions could focus on investigating Romania’s economic performance in comparison to other European states, providing an expanded perspective on economic convergence. Additionally, evaluating the model in which European cohesion funds contribute to reducing regional disparities could be another research directive.

Author Contributions

Conceptualization, I.N., I.G., C.D. and Ș.I.; methodology, I.G. and I.N.; software, I.G., I.N. and Ș.I.; validation, I.G., I.N., C.D., N.C. and Ș.I.; formal analysis, I.G. and I.N.; investigation, N.C. and Ș.I.; resources, I.G.; data curation, N.C. and Ș.I.; writing—original draft preparation, I.G. and I.N.; writing—review and editing, C.D., Ș.I. and N.C.; visualization, I.G., I.N., C.D., N.C. and Ș.I.; supervision, I.G., I.N., C.D., N.C. and Ș.I.; project administration, C.D. and I.N.; funding acquisition, C.D., I.N. and Ș.I. All authors have read and agreed to the published version of the manuscript.

Funding

This work was funded by the EU’s NextGenerationEU instrument through the National Recovery and Resilience Plan of Romania—Pillar III-C9-I8, managed by the Ministry of Research, Innovation and Digitization, within the project entitled “Place-based Economic Policy in EU’s Periphery—fundamental research in collaborative development and local resilience. Projections for Romania and Moldova (PEPER)”, contract No. 760045/23.05.2023, code CF 275/30.11.2022. This paper was co-financed by the Bucharest University of Economic Studies during the Ph.D. program.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The data presented in this study are available on request from the corresponding author.

Conflicts of Interest

The authors declare no conflicts of interest.

Abbreviations

AcronymDefinition
LLC panel unit root testLevin, Lin and Chu test
IPS testIm, Pesaran, and Shin test
ADFAugmented Dickey–Fuller
pARDLPanel Autoregressive Distributed Lag
ECTError Correction Term
PQRPanel Quantile Regression
OLSOrdinary Least Squares
GDPRRegional Gross Domestic Product
CTCRDResearch and Development Expenditure
ITIT Equipment in the Administration of Public and Private University Education Unit
FCLNumber of Universities
NMPSAverage Number of Employees

References

  1. Peng, B.; Sheng, X.; Wei, G. Does Environmental Protection Promote Economic Development? From the Perspective of Coupling Coordination between Environmental Protection and Economic Development. Environ. Sci. Pollut. Res. 2020, 27, 39135–39148. [Google Scholar] [CrossRef] [PubMed]
  2. Zakari, A.; Oluwaseyi Musibau, H. Sustainable Economic Development in OECD Countries: Does Energy Security Matter? Sustain. Dev. 2023, 32, 1337–1353. [Google Scholar] [CrossRef]
  3. Yang, Z.; Wang, H. Green Connection to the World: Building Regional Responsible Sustainable Development Strategies. Sustain. Dev. 2024, 32, 3555–3568. [Google Scholar] [CrossRef]
  4. Cai, J.; Chen, Q.; Zhang, Z. Balancing Environmental Sustainability and Economic Development: Perspectives from New Structural Economics. Sustainability 2024, 16, 1124. [Google Scholar] [CrossRef]
  5. Balan, M. Estimating Economic and Social Regional Disparities in Romania. Academia Brâncuși 2018, 3, 5–18. [Google Scholar]
  6. Russu, C. Aspects of Regional Development in Romania. Econ. Isights Trends Chall. 2014, III, 25–32. [Google Scholar]
  7. Goschin, Z. Regional Growth in Romania after Its Accession to EU: A Shift-Share Analysis Approach. Procedia Econ. Financ. 2014, 15, 169–175. [Google Scholar] [CrossRef]
  8. Otoiu, A.; Titan, E. Socio-Economic Convergence in the EU at National and Regional Level. Procedia Econ. Financ. 2015, 23, 1090–1095. [Google Scholar] [CrossRef]
  9. Pridham, G. Romania and EU Membership in Comparative Perspective: A Post-Accession Compliance Problem?—The Case of Political Conditionality. Perspect. Eur. Politics Soc. 2007, 8, 168–188. [Google Scholar] [CrossRef]
  10. Dornean, A. Main Factors and Implications of Policies Implemented in Romania in the Context of the Current Global Crisis. Procedia Soc. Behav. Sci. 2012, 58, 454–463. [Google Scholar] [CrossRef]
  11. Moussis, N. Guide to European Policies, 12th revised ed.; European Study Service: Brussels, Belgium, 2007; ISBN 978-2-930119-42-x. [Google Scholar]
  12. Gavriluţ, D.; Bădulescu, D.; University of Oradea, Romania; Jaganjac, J.; University “‘Vitez’” Bosnia and Herzegovina. The role of R&D sector on fostering economic growth. Insights from romania. In Proceedings of the 17th International Management Conference “Management Beyond Crisis: Rethinking Business Performance”, Bucharest, Romania, 2–3 November 2023. [Google Scholar]
  13. Kocsis, L.Z. The Importance of the Development of Industry in Romania and Its Development Potential. RETP 2024, 19, 56–76. [Google Scholar] [CrossRef]
  14. Androniceanu, A.; Georgescu, I.; Nica, E.; Popescu, G. A Computational Analysis of the Romanian Interregional Policy. In Proceedings of the 31st International Business Information Management Association Conference (IBIMA), Milan, Italy, 25–26 April 2018; pp. 1952–1962. [Google Scholar]
  15. Androniceanu, A.; Georgescu, I. Regional Policy in Romania: Main Changes and Dilemmas of Practice; London Scientific: London, UK, 2018; ISBN 978-0-9954618-5-7. [Google Scholar]
  16. Behrens, K.; Thisse, J.-F. Regional Economics: A New Economic Geography Perspective. Reg. Sci. Urban Econ. 2007, 37, 457–465. [Google Scholar] [CrossRef]
  17. Bailey, D.; Clark, J.; Colombelli, A.; Corradini, C.; De Propris, L.; Derudder, B.; Fratesi, U.; Fritsch, M.; Harrison, J.; Hatfield, M.; et al. Regions in a Time of Pandemic. Reg. Stud. 2020, 54, 1163–1174. [Google Scholar] [CrossRef]
  18. Martin, R.; Sunley, P. Regional Economic Resilience: Evolution and Evaluation. In Handbook on Regional Economic Resilience; Bristow, G., Healy, A., Eds.; Edward Elgar Publishing: Cheltenham, UK, 2020; ISBN 978-1-78536-086-2. [Google Scholar]
  19. Trippl, M.; Fastenrath, S.; Isaksen, A. Rethinking Regional Economic Resilience: Preconditions and Processes Shaping Transformative Resilience. Eur. Urban Reg. Stud. 2024, 31, 101–115. [Google Scholar] [CrossRef]
  20. Gumpert, M. Regional Economic Disparities under the Solow Model. Qual. Quant. 2024, 58, 1017–1037. [Google Scholar] [CrossRef]
  21. Liu, Y.; Yang, R.; Sun, M.; Zhang, L.; Li, X.; Meng, L.; Wang, Y.; Liu, Q. Regional Sustainable Development Strategy Based on the Coordination between Ecology and Economy: A Case Study of Sichuan Province, China. Ecol. Indic. 2022, 134, 108445. [Google Scholar] [CrossRef]
  22. Szopik-Depczyńska, K.; Cheba, K.; Bąk, I.; Kędzierska-Szczepaniak, A.; Szczepaniak, K.; Ioppolo, G. Innovation Level and Local Development of EU Regions. A New Assessment Approach. Land Use Policy 2020, 99, 104837. [Google Scholar] [CrossRef]
  23. Ferreira, C.; Broughton, K.; Broadhurst, K.; Ferreira, J. Collaborative Innovation in a Local Authority—‘Local Economic Development-by-Project’? Public Manag. Rev. 2024, 26, 1405–1423. [Google Scholar] [CrossRef]
  24. Ma, R.; Sun, J.; Yin, J. Research on the Impact Mechanism of Smart City Construction on Economic Growth—An Analysis Based on the Schumpeterian Innovation Theory Framework. Systems 2024, 12, 251. [Google Scholar] [CrossRef]
  25. Shi, Y.; Jiang, Y.; Xie, C.; Li, C. Regional Innovation and Sustainable Development Interplay: Analyzing the Spatial Externalities of Domestic Demand in the New Development Paradigm. Sustainability 2024, 16, 2365. [Google Scholar] [CrossRef]
  26. Pusz, M.; Jonas, A.E.G.; Deutz, P. Knitting Circular Ties: Empowering Networks for the Social Enterprise-Led Local Development of an Integrative Circular Economy. Circ. Econ. Sust. 2024, 4, 201–232. [Google Scholar] [CrossRef]
  27. Anda, M.I.; Gabriela, B.; Nicolae, E.; Rehman, A. Economic Growth Drivers in Romania: Evidence from a NARDL Analysis. Sustainability 2023, 15, 5916. [Google Scholar] [CrossRef]
  28. Chiriță, N.; Ciurea, C.; Nica, I.; Delcea, C. Exploring Academic Networks in the Fields of Cybernetics and Economics: A Bibliometric Analysis of the Economic Computation and Economic Cybernetics Studies and Research Journal. ECECSR 2024, 58, 307–322. [Google Scholar] [CrossRef]
  29. Gurgu, E.; Frăsineanu, D.; Simbotin, G. How Creative Sectors Can Contribute to the Sustainable Developemnt of Romania. Ann. Spiru Haret Univ. Econ. Ser. 2024, 24, 1–27. [Google Scholar]
  30. Deng, P.; Lu, S.; Xiao, H. Evaluation of the Relevance Measure between Ports and Regional Economy Using Structural Equation Modeling. Transp. Policy 2013, 27, 123–133. [Google Scholar] [CrossRef]
  31. Park, J.S.; Seo, Y.-J. The Impact of Seaports on the Regional Economies in South Korea: Panel Evidence from the Augmented Solow Model. Transp. Res. Part E Logist. Transp. Rev. 2016, 85, 107–119. [Google Scholar] [CrossRef]
  32. Chen, Z. Measuring the Regional Economic Impacts of High-Speed Rail Using a Dynamic SCGE Model: The Case of China. Eur. Plan. Stud. 2019, 27, 483–512. [Google Scholar] [CrossRef]
  33. Kauppila, P.; Karjalainen, T.P. A Process Model to Assess the Regional Economic Impacts of Fishing Tourism: A Case Study in Northern Finland. Fish. Res. 2012, 127–128, 88–97. [Google Scholar] [CrossRef]
  34. Madsen, B.; Jensen-Butler, C. Theoretical and Operational Issues in Sub-Regional Economic Modelling, Illustrated through the Development and Application of the LINE Model. Econ. Model. 2004, 21, 471–508. [Google Scholar] [CrossRef]
  35. Charalampidis, I.; Karkatsoulis, P.; Capros, P. A Regional Economy-Energy-Transport Model of the EU for Assessing Decarbonization in Transport. Energies 2019, 12, 3128. [Google Scholar] [CrossRef]
  36. Mushtaq, S.; Cockfield, G.; White, N.; Jakeman, G. Modelling Interactions between Farm-Level Structural Adjustment and a Regional Economy: A Case of the Australian Rice Industry. Agric. Syst. 2014, 123, 34–42. [Google Scholar] [CrossRef]
  37. Garidzirai, R.; Muzindutsi, P.-F. A Panel ARDL Analsis of the Productivity of Key Economic Sectors Contributing to Local Economic Growth in an Emerging Country. Stud. Univ. Babes-Bolyai Oeconomica 2020, 65, 39–53. [Google Scholar] [CrossRef]
  38. Sadik-Zada, E.R.; Loewenstein, W.; Hasanli, Y. Production Linkages and Dynamic Fiscal Employment Effects of the Extractive Industries: Input-Output and Nonlinear ARDL Analyses of Azerbaijani Economy. Miner. Econ. 2021, 34, 3–18. [Google Scholar] [CrossRef]
  39. Qalandarovna Abdurakhmanova, G.; Jeong, J.Y.; Rakhimjon Ugli Oqmullayev, R.; Utkir Ugli Karimov, M. The Impact of Tourism on Employment and Economic Growth in Uzbekistan: An ARDL Bounds Testing Approach. In Proceedings of the 5th International Conference on Future Networks & Distributed Systems, Dubai, United Arab Emirates, 15–16 December 2021; ACM: New York, NY, USA, 2021; pp. 431–439. [Google Scholar]
  40. Levin, A.; Lin, C.-F.; James Chu, C.-S. Unit Root Tests in Panel Data: Asymptotic and Finite-Sample Properties. J. Econom. 2002, 108, 1–24. [Google Scholar] [CrossRef]
  41. Im, K.S.; Pesaran, M.H.; Shin, Y. Testing for Unit Roots in Heterogeneous Panels. J. Econom. 2003, 115, 53–74. [Google Scholar] [CrossRef]
  42. Maddala, G.S.; Wu, S. A Comparative Study of Unit Root Tests with Panel Data and a New Simple Test. Oxf. Bull. Econ. Stat. 1999, 61, 631–652. [Google Scholar] [CrossRef]
  43. Payne, J.E.; Lee, J.; Islam, M.T.; Nazlioglu, S. Stochastic Convergence of per Capita Greenhouse Gas Emissions: New Unit Root Tests with Breaks and a Factor Structure. Energy Econ. 2022, 113, 106201. [Google Scholar] [CrossRef]
  44. Babar, H.; Hamid, K.; Rasheed, M.; Rasool, F.; Saeed, M.Y. Do Macroeconomic Factors Are Interlinked with Stock Markets in Asian Emerging Economies. A Panel Data Approach. Int. J. Bus. Econ. Aff. 2024, 9, 42–51. [Google Scholar] [CrossRef]
  45. Kiran Baygin, B.; Çil, N. Persistence of Shocks on Non-Renewable and Renewable Energy Consumption: Evidence from 15 Leading Countries with Fourier Unit Root Test. Environ. Dev. Sustain. 2023, 26, 5915–5934. [Google Scholar] [CrossRef]
  46. Kilinc-Ata, N.; Barut, A.; Citil, M. Do Military Expenditures Have an Impact on the Adoption of Renewable Energy in OECD Nations? Evidence from a Panel Cointegration Test Approach. IJESM 2024. [Google Scholar] [CrossRef]
  47. Kathuria, K.; Kumar, N. Examining the Dynamics of India’s Major Exchange Rates Using Fourier Nonlinear Quantile Unit Root Test. Asia-Pac. Financ. Markets 2024, 1–16. [Google Scholar] [CrossRef]
  48. Dickey, D.A.; Fuller, W.A. Distribution of the Estimators for Autoregressive Time Series with a Unit Root. Journal of the American Statistical Association 1979, 74, 427–431. [Google Scholar] [CrossRef]
  49. Choi, I. Unit Root Tests for Panel Data. J. Int. Money Financ. 2001, 20, 249–272. [Google Scholar] [CrossRef]
  50. Kaur, N.; Singh, B. Dynamics of Resource Allocation–Firm Performance Relationship: Panel Autoregressive Distributed Lag Approach. Manag. Decis. Econ. 2024, 45, 3663–3676. [Google Scholar] [CrossRef]
  51. Epo, B.N.; Tapche Ndam, Y.F.; Abiala, A.M. Knowledge Economy and Financial Development in Developing Countries: Evidence from a Panel Autoregressive Distributed-Lag (ARDL) Approach. J. Knowl. Econ. 2024, 1–55. [Google Scholar] [CrossRef]
  52. Ghazouani, T.; Maktouf, S. Impact of Natural Resources, Trade Openness, and Economic Growth on CO2 Emissions in Oil-exporting Countries: A Panel Autoregressive Distributed Lag Analysis. In Natural Resources Forum; Blackwell Publishing Ltd: Oxford, UK, 2023. [Google Scholar] [CrossRef]
  53. Ben Abdallah, A. The Relationship Between Trade Openness, Foreign Direct Investment Inflows, and Economic Growth in Middle East and North of Africa Region: Autoregressive Distributed Lag Model vs. Vector Error Correction Model. J. Knowl. Econ. 2024, 15, 1118–1141. [Google Scholar] [CrossRef]
  54. Pesaran, M.H.; Shin, Y. Cointegration and Speed of Convergence to Equilibrium. J. Econom. 1996, 71, 117–143. [Google Scholar] [CrossRef]
  55. Zardoub, A. Exploring the Links between Financial Flows and Economic Growth: A Panel ARDL Approach. PRR 2023, 7, 90–104. [Google Scholar] [CrossRef]
  56. Hidthiir, M.H.B.; Ahmad, Z.; Junoh, M.Z.M.; Yusof, M.F.B. Dynamics of Economic Growth in ASEAN-5 Countries: A Panel ARDL Approach. Discov. Sustain. 2024, 5, 145. [Google Scholar] [CrossRef]
  57. Sek, S.K.; Sim, K.Y.; Mukherjee, D. Panel ARDL and NARDL Models: Role of Agriculture in Economic Growth in Selected Asian Countries. In Applied Econometric Analysis Using Cross Section and Panel Data; Mukherjee, D., Ed.; Contributions to Economics; Springer Nature Singapore: Singapore, 2023; pp. 471–503. ISBN 978-981-9949-01-4. [Google Scholar]
  58. Mehmood, K.; Tauseef Hassan, S.; Qiu, X.; Ali, S. Comparative Analysis of CO2 Emissions and Economic Performance in the United States and China: Navigating Sustainable Development in the Climate Change Era. Geosci. Front. 2024, 15, 101843. [Google Scholar] [CrossRef]
  59. Koenker, R.; Hallock, K.F. Quantile Regression. J. Econ. Perspect. 2001, 15, 143–156. [Google Scholar] [CrossRef]
  60. Baffour Gyau, E.; Li, Y.; Adu, D. Investigating the Impact of ICT on Transport-Based CO2 Emissions: Empirical Evidence from a Quantile Cointegration Regression Analysis. Environ. Sci. Pollut. Res. 2023, 31, 4606–4629. [Google Scholar] [CrossRef] [PubMed]
  61. Işık, C.; Bulut, U.; Ongan, S.; Islam, H.; Irfan, M. Exploring How Economic Growth, Renewable Energy, Internet Usage, and Mineral Rents Influence CO2 Emissions: A Panel Quantile Regression Analysis for 27 OECD Countries. Resour. Policy 2024, 92, 105025. [Google Scholar] [CrossRef]
  62. Elmonshid, L.B.E.; Sayed, O.A.; Awad Yousif, G.M.; Eldaw, K.E.H.I.; Hussein, M.A. The Impact of Financial Efficiency and Renewable Energy Consumption on CO2 Emission Reduction in GCC Economies: A Panel Data Quantile Regression Approach. Sustainability 2024, 16, 6242. [Google Scholar] [CrossRef]
  63. Granger, C.W.J. Investigating Causal Relations by Econometric Models and Cross-Spectral Methods. Econometrica 1969, 37, 424–438. [Google Scholar] [CrossRef]
  64. National Institute of Statistics Statistical Data. Available online: https://insse.ro/cms/en (accessed on 9 October 2024).
  65. Ministry of Finance. 2003 Regular Report on Romania’s Progress Towards Accession. Available online: https://mfinante.gov.ro/documents/35673/228074/regular_report2003.pdf (accessed on 15 October 2024).
Figure 1. The methodological steps applied in the analysis of regional and local economic performance in Romania.
Figure 1. The methodological steps applied in the analysis of regional and local economic performance in Romania.
Sustainability 16 09287 g001
Figure 2. Multidimensional analysis of correlations between economic variables: Pearson, Spearman, Kendall, and Covariance.
Figure 2. Multidimensional analysis of correlations between economic variables: Pearson, Spearman, Kendall, and Covariance.
Sustainability 16 09287 g002
Figure 3. GDPR Evolution by Region.
Figure 3. GDPR Evolution by Region.
Sustainability 16 09287 g003
Figure 4. CTCRD evolution by region.
Figure 4. CTCRD evolution by region.
Sustainability 16 09287 g004
Figure 5. IT Evolution by Region.
Figure 5. IT Evolution by Region.
Sustainability 16 09287 g005
Figure 6. FCL evolution by region.
Figure 6. FCL evolution by region.
Sustainability 16 09287 g006
Table 1. Summary of Key Studies on Regional and Local Economic Impacts.
Table 1. Summary of Key Studies on Regional and Local Economic Impacts.
First Author; Year; Journal; ReferenceScopeMethodologyKey FindingsRelevance to Regional and Local Economy
Deng, P.; 2013; Transport Policy; [30]This study explores the relationship between ports (port demand and supply, port value-added activity) and the regional economy from a logistics perspective, providing decision support for decision makers.Structural equation modeling (SEM) approach using samples from the five coastal port clusters in China.The results showed that port supply positively influences port demand and that port value-added activities have a positive impact on the regional economy; however, port demand and supply did not have significantly positive effect on the regional economy.The study highlights the impact of port activities on the regional economy and provides directions for decision makers in port development planning.
Park, J.S.; 2016; Transportation Research Part E: Logistics and Transportation Review; [31]The study investigates the economic impact of seaports on regions in Korea using an econometric analysis based on the extended Solow model.Econometric analysis based on panel data covering all regions in Korea between 2000 and 2013.Econometric analysis shows that cargo ports without sufficient traffic hinder regional economic growth while cargo ports contribute to regional economic growth only when traffic is sufficient. Container port activities have a positive impact on regional economic growth and investment in ports indirectly leads to economic growth.The study provides a clear understanding of the role of ports in the economic growth of Korea’s regions, suggesting the importance of adequate port traffic and indirect investment in ports.
Chen, Z.; 2019; European Planning Studies; [32]The study analyzes the regional economic impact of high-speed rail (HSR) infrastructure development in China using a dynamic and spatial computable general equilibrium framework.Dynamic and spatial computable general equilibrium modeling framework based on real data from 2002 to 2013.The development of HSR infrastructure in China has had a positive regional economic impact. Real GDP growth boosted by rail infrastructure investment was significant in the South-West but lower in the developed eastern regions. Contributions to economic growth come mainly from increased productivity in the rail transport sector and the stimulating effect of rail infrastructure investments.The study provides insights into the impact of HSR infrastructure on regional economies and suggests that infrastructure investment can spur economic growth, particularly in less developed regions.
Kauppila, P.; 2012; Fisheries Research; [33]The study analyzes the regional economic impact of recreational fishing tourism, with an emphasis on the development of a procedural model applicable in rural areas.Development of a procedural model based on existing statistics to assess regional economic impact, with practical applicability and low data collection costs.The proposed model calculates regional economic indicators for fishing tourism, being applied in the area of the Iijoki river in Finland. Although it ignores multiplicative effects, it provides a simple and efficient method of analysis.The study contributes to the understanding of the impact of fishing tourism on the local and regional economy, providing a practical tool for planners and researchers, with the potential to revitalize rural areas.
Madsen, B.; 2004; Economic Modelling; [34]The study analyzes the links between the regional economy and activities at the sub-regional level, focusing on disaggregated general equilibrium models applied to transport costs and bridge tolls in Denmark.Using the sub-regional disaggregated general equilibrium model, called LINE, to assess the impact of changes in transport costs on economic activity.The study highlights the fact that traditional economic models at the regional level fail to capture the fundamental economic mechanisms at the sub-regional level. The application of the model shows how changes in transport costs influence economic activity at the sub-regional level in Denmark.The relevance of this study lies in clarifying how sub-regional economic changes, such as transport costs, can influence the regional economy, providing a better understanding of economic linkages at different spatial levels.
Charalampidis, I.; 2019; Energies; [35]The study presents a new macroeconomic–regional model (GEM-E3-R) for assessing the impact of transport sector restructuring on regional economies in the EU in the context of the EU’s decarbonization strategy.A computable general equilibrium (CGE) model with two layers: one at the national level (multi-sectoral, multi-country) and one at the regional level, which assesses the impact of the restructuring of the transport sector using a dynamic agglomeration–dispersion mechanism and a gravity model for inter-regional flows.The restructuring of the transport sector, in particular the electrification of car mobility and the use of advanced biofuels, will generate positive economic effects on regional economies in the EU, with differentiations depending on the industrial structure and the ability of regions to adopt new technologies.The study highlights the importance of infrastructure investment and new technologies in the transport sector for regional development, providing a rigorous assessment of the long-term economic impact of the EU’s decarbonization strategy.
Mushtaq, S.; 2014; Agricultural Systems; [36]The study investigates farm-level structural adjustments and their impact on regional economies in Australia in the context of climate change and water use policy reforms, with a focus on the rice industry.Integrating empirical evidence on farm-level structural adjustments with a regional computable general equilibrium (CGE) model to assess the sectoral and regional impacts of climate change and water use policy.The study shows that the existing diversity in agricultural production has allowed adaptations at the farm level, but additional water loss and the adoption of less intensive production methods will reduce the net number of agricultural businesses, which will affect rice production at the regional level. The negative impacts on the regional economy are partially offset by the redistribution of resources at the national level. Relocating rice production to the north would require strong government support.The study highlights the impact of water use policies and climate change on regional economies, highlighting the need for government support to adapt the rice industry to the new conditions.
Garidzirai, R; 2020; Studia Universitatis Babes-Bolyai Oeconomica; [37]The research examines the principal industries that might foster local economic development in South Africa, notwithstanding the general economic stagnation.A Panel-Distributed Autoregressive Model (ARDL) using PMG, MG, and DFE estimators on annual data from 1996 to 2015 to analyze productivity effects in six sectors.Productivity in the construction, transport, trade, manufacturing and finance sectors positively influences long-term economic growth, while productivity in mining and tourism has a negative effect. In the short term, all sectors, except trade and transport, contribute positively to local economic growth.The study identifies sectors that can boost local economic growth in South Africa and recommends investment in infrastructure and skills development in mining and tourism to boost regional productivity.
Sadik-Zada, E.R.; Mineral Economics; [38]The study analyzes the production linkages and employment effects of the oil sector on the rest of the Azerbaijani economy, using input–output tables for the years 2006, 2008 and 2009.Input–output analysis of output and job creation multipliers followed by a nonlinear distributed autoregressive (ARDL) model for assessing the long-run effects of oil revenues on employment.Despite advanced infrastructure and local content policies, the integration of the oil sector into the domestic economy is weak, and production and job creation multipliers have declined slightly after 3 years of exponential growth in oil production. Investments in the processing, construction, and network industries have the largest output linkages, and agriculture, education, health and the public sector have the largest job creation effect.The study highlights the impact of oil revenues on job creation in Azerbaijan and suggests the need for additional investment in oil-related industries to boost regional manufacturing linkages.
Abdurakhmanova, G.Q.; 2021; Proceedings of the 5th International Conference on Future Networks and Distributed Systems; [39]The study investigates the influence of tourism on employment and economic growth in Uzbekistan using empirical data from 1991 to 2020.An autoregressive distributed model (ARDL) to analyze the long- and short-run connections between global tourism arrivals, employment, and real GDP.The results show a close relationship between tourism, employment, and economic growth in Uzbekistan, with a unidirectional causality from tourism to economic growth. The increase in the number of international tourists has a positive impact on the economy and employment.The study highlights the strategic role that tourism can play in boosting regional economic growth and employment, suggesting that Uzbekistan could achieve economic prosperity by supporting the tourism sector.
Table 2. Descriptive statistics.
Table 2. Descriptive statistics.
GDPRCTCRDFCLITNMPS
Mean11.1512.404.229.0313.13
Median11.1612.344.179.0713.14
Maximum12.7115.535.2810.5513.71
Minimum9.7310.653.407.4512.72
Std. Dev.0.591.070.480.830.22
Skewness0.050.750.250.110.65
Kurtosis3.043.202.481.853.50
Jarque–Bera0.0815.683.489.0112.98
Probability0.950.000.170.010.00
Table 3. Panel unit root tests.
Table 3. Panel unit root tests.
At Levels
GDPRCTCRDFCLITNMPS
Unit root (Common Unit Root Process)
LLC−5.41 ***
(0.00)
1.68
(0.92)
−7.16 ***
(0.00)
−5.60 ***
(0.00)
−1.43 *
(0.07)
Unit root (Individual Root Process)
IPS−2.02 **
(0.02)
−1.72 **
(0.04)
−3.28 ***
(0.00)
−4.65 ***
(0.00)
−1.07
(0.14)
ADF Fisher Chi-Square−26.18 *
(0.05)
43.34 ***
(0.00)
40.18 ***
(0.00)
58.73 ***
(0.00)
18.33
(0.30)
At first difference
Unit root (Common Unit Root Process)
LLC−5.38 ***
(0.00)
−9.17 ***
(0.00)
−10.54 ***
(0.00)
−6.84 ***
(0.00)
−4.78 ***
(0.00)
Unit root (Individual Root Process)
IPS−4.07 ***
(0.00)
−9.06 ***
(0.00)
−7.84 ***
(0.00)
−6.65 ***
(0.00)
−4.73 ***
(0.00)
ADF-Fisher Chi-Square43.69 ***
(0.00)
98.68 ***
(0.00)
82.53 ***
(0.00)
70.87 ***
(0.00)
50.10 ***
(0.00)
*, **, *** significant at 10%, 5% and 1% level.
Table 4. Panel ARDL (2,2,2,2,2).
Table 4. Panel ARDL (2,2,2,2,2).
VariableCoefficientStd. Errort-StatisticProb.
Long-run equation
CTCRD0.050.041.380.17
FCL−1.090.26−4.170.00 ***
IT0.530.095.430.00 ***
NMPS4.400.2417.900.00 ***
Short-run equation
COINTEQ01−0.320.04−6.730.00 ***
D (GDPR (−1))−0.360.16−2.110.03 **
D(CTCRD)−0.030.01−2.390.01 ***
D (CTCRD (−1))−0.040.01−2.850.00 ***
D(FCL)−0.070.13−0.570.56
D (FCL (−1))−0.280.08−3.240.00 ***
D(IT)−0.260.09−2.800.00 ***
D (IT (−1))−0.120.09−1.350.18
D(NMPS)0.420.202.020.04 **
D (NMPS (−1))−0.250.23−1.100.27
C−15.552.37−6.540.00 ***
Notes: **, *** significant at 5% and 1% level.
Table 5. Panel quartile regression—25th quartile.
Table 5. Panel quartile regression—25th quartile.
VariableCoefficientStd. Errort-StatisticProb.
CTCRD0.080.061.330.18
FCL−1.090.12−8.750.00 ***
IT0.690.0414.060.00 ***
NMPS0.620.0511.380.00 ***
Pseudo R-squared0.53Mean dependent var 11.15
Adjusted R-squared0.52S.D. dependent var 0.59
S.E. of regression0.38Objective 14.19
Quantile dependent var10.79Restr. objective 30.69
Sparsity0.79
Notes: *** significant at 1% level.
Table 6. Panel quartile regression—median.
Table 6. Panel quartile regression—median.
VariableCoefficientStd. Errort-StatisticProb.
CTCRD0.180.053.150.00 ***
FCL−0.840.09−8.520.00 ***
IT0.520.076.920.00 ***
NMPS0.570.0413.830.00 ***
Pseudo R-squared0.50Mean dependent var 11.15
Adjusted R-squared0.49S.D. dependent var 0.59
S.E. of regression0.31Objective 18.25
Quantile dependent var11.15Restr. objective 36.65
Sparsity0.58
Notes: *** significant at 1% level.
Table 7. Panel quartile regression—75th quartile.
Table 7. Panel quartile regression—75th quartile.
VariableCoefficientStd. Errort-StatisticProb.
CTCRD0.270.064.190.00 ***
FCL−0.730.16−4.360.00 ***
IT0.410.094.250.00 ***
NMPS0.530.059.900.00 ***
Pseudo R-squared0.48Mean dependent var 11.15
Adjusted R-squared0.47S.D. dependent var 0.59
S.E. of regression0.38Objective 15.18
Quantile dependent var11.53Restr. objective 29.59
Sparsity0.84
Notes: *** significant at 1% level.
Table 8. Biased corrected linear dynamic approach—Robustness check.
Table 8. Biased corrected linear dynamic approach—Robustness check.
VariableCoefficientStd. Errorz-StatisticProb.[95% Confidence Interval]
CTCRD0.010.010.190.00 ***[−0.04, 0.04]
FCL−0.310.06−2.960.00 ***[−0.52, −0.10]
IT−0.020.02−0.580.26[−0.08, 0.04]
NMPS0.500.0516.000.00 ***[0.85, 1.09]
Constant−3.140.68−8.040.00 ***[−10.99, −6.68]
*** significant at 1% level.
Table 9. Granger panel causality test.
Table 9. Granger panel causality test.
Null HypothesisF-StatisticProb.Conclusion
CTCRD does not Granger cause GDPR17.20 6 × 10 5 *** C T C R D   G D P R
GDPR does not Granger cause CTCRD32.04 8 × 10 8 *** G D P R   C T C R D
FCL does not Granger cause GDPR22.86 4 × 10 6 *** F C L   G D P R
GDPR does not Granger cause FCL3.540.06 ** G D P R   F C L
IT does not Granger cause GDPR2.150.14
GDPR does not Granger cause IT7.870.00 *** G D P R   I T
NMPS does not Granger cause GDPR38.15 6 × 10 9 *** N M P S   G D P R
GDPR does not Granger cause NMPS5.910.01 *** G D P R   N M P S
FCL does not Granger cause CTCRD2.840.09 * F C L   C T C R D
CTCRD does not Granger cause FCL14.840.00 *** C T C R D   F C L
IT does not Granger cause CTCRD19.08 2 × 10 5 *** I T   C T C R D
CTCRD does not Granger cause IT1.080.29
NMPS does not Granger cause CTCRD10.410.00 *** N M P S   C T C R D
CTCRD does not Granger cause NMPS0.730.39
IT does not Granger cause FCL7.270.00 *** I T   F C L
FCL does not Granger cause IT15.500.00 *** F C L   I T
NMPS does not Granger cause FCL2.130.14
FCL does not Granger cause NMPS1.890.17
NMPS does not Granger cause IT2.490.12
IT does not Granger cause NMPS2.530.11
*, **, *** significant at 10%, 5%, and 1% level.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Georgescu, I.; Nica, I.; Delcea, C.; Chiriță, N.; Ionescu, Ș. Assessing Regional Economic Performance in Romania Through Panel ARDL and Panel Quantile Regression Models. Sustainability 2024, 16, 9287. https://doi.org/10.3390/su16219287

AMA Style

Georgescu I, Nica I, Delcea C, Chiriță N, Ionescu Ș. Assessing Regional Economic Performance in Romania Through Panel ARDL and Panel Quantile Regression Models. Sustainability. 2024; 16(21):9287. https://doi.org/10.3390/su16219287

Chicago/Turabian Style

Georgescu, Irina, Ionuț Nica, Camelia Delcea, Nora Chiriță, and Ștefan Ionescu. 2024. "Assessing Regional Economic Performance in Romania Through Panel ARDL and Panel Quantile Regression Models" Sustainability 16, no. 21: 9287. https://doi.org/10.3390/su16219287

APA Style

Georgescu, I., Nica, I., Delcea, C., Chiriță, N., & Ionescu, Ș. (2024). Assessing Regional Economic Performance in Romania Through Panel ARDL and Panel Quantile Regression Models. Sustainability, 16(21), 9287. https://doi.org/10.3390/su16219287

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop