New insights on immigration, fiscal policy and unemployment rate in EU countries – A quantile regression approach

Abstract

Free movement of production factors in the enlarged EU has led to immigration flows from East to West and from South to North with a significant impact on EU labor markets. Fiscal federalism also determined large immigration flows into EU area and affected unemployment rate in the EU countries. The aim of this paper is to investigate the impact of factors such as number of immigrants, tax on profits, social contributions, economic growth and population growth on unemployment rates in EU area using a panel quantile regression and an PMG-ARDL approach as robustness test during 1991–2020. The results show a positive association between population growth and unemployment rate, whereas the remaining exogenous factors are negatively associated with unemployment rate. Still, social contributions are statistically significant only for upper quantiles. The overall impact of social contribution on unemployment rate is positive as per PMG-ARDL estimations. We have also demonstrated that the immigrant flows impact on unemployment rate is very weak. The factors that are exerting the most significance influence on unemployment rate, are economic and population growth, followed by tax on profits. Findings support policy recommendation in EU area in terms of fiscal policy.

1. Introduction

Migration has persisted as a recurring phenomenon throughout human history. In today's interconnected global context, international migration has taken on a significant role as a direct response to intricate socio-economic factors that span diverse geographical regions. Previous research has acknowledged the considerable economic consequences brought about by immigration. Nevertheless, a consensus regarding the exact impact of immigration on the labor market remains elusive. Economic impact of immigration is a complex interplay of multiple factors. Each European country's unique combination of immigration policies, economic conditions, labor market dynamics, and societal attitudes contributes to a distinctive experience with immigration's economic effects. Previous works [1] highlighted on the effects of immigration on economic growth. He delved into how the outcomes resulting from immigration are not uniform; instead, they vary depending on specific time periods and geographic locations. This variability in outcomes can result in both positive and negative impacts on economic conditions. However, it's essential to recognize that the worldwide economic recession stemming from the 2008 financial crisis, or pandemic crisis, temporarily led to a reduction in cross-border movement of people. This effect emerged because the movement of labor tends to follow economic cycles, where during economic downturns, the mobility of workers decreases. This phenomenon is termed “pro-cyclical,” indicating that labor mobility tends to move in the same direction as economic cycles. As a result, the 2008 financial crisis caused a temporary slowdown in the movement of people across borders due to the unfavorable economic conditions prevailing at that time [2].

By evidence that immigrants are often more affected when the economy is not doing well. This is because they might have fewer stable jobs and lower positions at work. Past studies have also found that discrimination against immigrants gets worse when fewer jobs are available. Also, when fewer people are needed for work, knowing people and having connections become even more important for getting a job. Immigrants usually have fewer of these connections compared to people from that place. This can make it even harder for them to find a job when the economy is not good. So, to sum it up, immigrants can have a tougher time when the economy is bad because their jobs are not as secure, they might be treated unfairly, and they might not know as many people to help them find work. The phenomenon of unemployment directly exerts an influence on individuals' capacity to acquire essential resources, exercise preferences in consumption, and uphold a specific standard of living. This implies that when individuals experience unemployment, there is a noteworthy effect on their capability to secure vital resources, make decisions regarding their consumption patterns, and sustain a particular quality of life [3,4]. Unemployment widens income inequality, hinders economic growth via wasted resources and reduced spending, links to poverty by denying basic needs, triggers migration for job prospects, and fuels economic instability. The lack of steady income for a significant portion of the population due to unemployment can also contribute to economic instability, disrupting consumer spending, tax revenues, and fiscal planning. Thus, addressing unemployment becomes essential for ensuring a balanced and thriving economy [3,5].

In the contemporary global context, the impact of international migration extends to a majority of nations, manifesting in various roles: some countries serve as sources of emigrants, others as destinations for immigrants, and frequently, many countries fulfill both functions simultaneously. The profiles of migrants often diverge markedly from those of the host populations, encompassing distinctions in terms of demographics, cultural attributes, and socio-economic attributes. This contrast highlights the diversity that migration introduces within societies. Furthermore, the pattern of migrant settlement is remarkably prevalent across the world, characterized by a strong tendency to concentrate in specific regions that hold particular appeal and in metropolitan areas that offer opportunities and resources. These locations, often referred to as attraction regions, draw migrants due to the prospect of better livelihoods, economic prospects, and the presence of established communities.

In essence, international migration has become a pervasive phenomenon, impacting countries in a variety of ways – as sources of emigrants seeking opportunities abroad, as destinations for immigrants enriching their societies, or sometimes as both simultaneously. This movement of people brings about a mix of cultural, demographic, and socio-economic dynamics that reshape the fabric of societies, with focal points of settlement centered in regions of high attraction and urban agglomerations.

In numerous advanced economies, the proportion of their populations comprising individuals born in foreign countries has surged to reach double-digit figures. This notable increase signifies the growing impact of international migration on these countries. During the Covid-19 pandemic, the employment experiences of immigrants have shown distinct patterns. There was a notably sharp decline in the early stages of the pandemic, followed by a remarkably robust increase in 2021. This rebound has restored immigrant employment rates to their pre-pandemic levels in OECD-Europe and Canada. However, the same recovery has not been seen in the United States. Simultaneously, owing to their prevalence in roles linked to economic cycles, immigrants also tend to be among the initial beneficiaries when the economy picks up. This trend is in line with previous findings by the OECD in 2019. Consider the hospitality sector in European OECD countries as an illustration. Here, over 25 % of employees originate from foreign countries, which is twice their representation in the broader job market. This disproportion is even more pronounced among recent arrivals. A similar pattern of significant overrepresentation can also be witnessed in Australia, Canada, Japan, New Zealand, and the United States [6].

Micro, small and medium-sized enterprises (SMEs) encompassing 99 % of the companies operating within the EU. These SMEs play a pivotal role as they are responsible for generating two-thirds of the jobs within the private sector. Furthermore, they contribute to over half of the total added value produced by businesses in the EU. This highlights the substantial impact of SMEs on job creation and economic value, underscoring their vital position within the European Union's business landscape. The SMEs employ around 100 million individuals, making them a critical contributor to the entrepreneurial drive and innovation within the EU. This contribution is vital for enhancing the competitiveness of companies in the European Union.1 Performance of Small and Medium Enterprises is also influenced by the tax structure of a country. The tax policies and rates can affect the cost of doing business, the profitability of SMEs, and their employment capacity. The taxes that SMEs are required to pay are often perceived as hindrances to enhancing their performance [7].

The unemployment rate was 8.5 percent of the total labor force in the European Union and progressively increased, peaking at 11.5 % in 1994. From 1995 to 2001, the rate fluctuated within a stable range with minor deviations. The year 2008 witnessed a significant decline to 7.2 %, likely due to a period of economic stability before the global financial crisis. However, in 2009, it surged to 9.1 % due to the crisis causing widespread economic downturns and job losses. The following decade, from 2009 to 2019, saw a consistent decline to 6.7 %, signifying EU economic recovery. In 2020, the rate slightly rose to 7.1 % due to COVID-19 disruptions. The relatively steady rates of 7 % in 2021 and 2022 suggest a phase of recovery and stabilization following the pandemic's initial impact [8].

A study [9] demonstrated that unemployment can be mitigated by tax cuts and increasing public production, which will lead to an increased public debt. They stated that this fiscal and budgetary policy can be efficient in addressing unemployment issues.

Another study [10] also analyzed Keynesian School and the New Classical School for achieving macroeconomic equilibrium. New Classical School stated that unemployment can't be reduced without sufficient accumulation of capital and that unemployment is transitory, while Keynesian School stated that full employment is not possible because of market imperfections. However, the authors claimed for a greater role of the expansionary fiscal policy. Policy makers should support micro and macro adjustments for effective fiscal policy in order to reduce unemployment rate.

Some authors [11] proved that fiscal expansion supports real wages increase and that increasing public spending actually increases unemployment through rising labor force participation into the labor market. Others [12] investigated the relation of fiscal policy with unemployment and stated that fiscal policy can be effective in overcoming the crises period and in mitigating unemployment issues, but only in the context of sticky prices. If the prices would be more flexible, the costly fiscal policy won't be necessary in fighting against unemployment. Another study [13] built a neo-classical model to investigate the effect of fiscal policy on outcome and unemployment multipliers in US. He proved the efficiency of using fiscal policy and public spendings on increasing GDP output and decreasing unemployment.

This study aims to investigate the nexus between immigration flows and unemployment rate in EU countries. It considers important aspects of fiscal policy such as tax on corporate profits and social contributions, because EU is a federal union in terms of fiscal policy, while it operates as an integrated monetary union. This is one of the main critics of the EU as optimal currency area, namely the fact that it displays large differences in terms of fiscal policies, while there is a single monetary policy. The fact that EU countries applies their own fiscal policy allows them to react differently to external shocks in case of crisis. Since free movement of production factors inside EU area has generated large migration flows from East to West and from South to North of Europe, so the impact of migration flows on the trend on unemployment rates in EU area is a topic which deserves a great attention. This investigation allows to prove if the federalism of fiscal policy exerts a significant impact on unemployment in the EU area and which is the immigration impact among EU countries on unemployment rates. This can provide some important policy recommendations for authorities who can adapt their fiscal measures in order to reduce unemployment and control the impact of immigration on the labor markets. We have investigated the nexus between immigration, tax on profits, social contributions and unemployment by applying 2^ndgenerations techniques such as panel quantile regression, for the robustness of the estimation results in the frame of cross-sectional dependence among the panel and we have included economic growth as GDP per capita and population growth as control variables into the estimation model.

2. Literature review

2.1. Economic growth and unemployment

The boom of the economic activity is responsible with creating new job opportunities. Still, this relation depends on many other factors as well and the exact stage of economic growth. Some authors [14] examined the connection between economic development and the unemployment rate in Slovakia during a phase of stabilization and economic growth. They found that the expected narrow, indirect correlation between GDP growth and the unemployment rate was not present. Consequently, the study concluded that the assumption that economic growth automatically resolves unemployment.

The relationship between economic growth and the unemployment rate was investigated for Albania, using Okun's law, which suggests that a 1 % decrease in unemployment results in a 3 % increase in GDP. The study covered the years from 2000 to 2013 and employed a straightforward regression methodology to analyze the correlation between GDP and the unemployment rate [15].

A panel analysis [16] investigated the relationship between unemployment and economic growth in 12 selected Asian countries spanning from 1982 to 2011. The research unveiled that a higher unemployment rate exerted a substantial negative impact on GDP per capita growth. Consequently, the study indicated that reducing unemployment constituted a more favorable strategy for attaining sustained economic growth and elevating the well-being of the population. Furthermore, they discovered that economic growth was significantly influenced by conventional factors such as inflation, population growth, gross capital formation, and trade openness. These determinants played pivotal roles in shaping the economic growth trajectory of the analyzed Asian countries.

A study performed for EU countries [17] investigated the relationship between economic growth and gender-specific unemployment rates in the European Union (EU). The research involved estimating Okun's coefficients for all EU countries, as well as specific groups of countries with similar characteristics. The findings revealed that male unemployment was more responsive to fluctuations in GDP compared to female unemployment. Additionally, the impact of changes in GDP on unemployment was more pronounced in EU countries with lower economic performance. These results shed light on the differing dynamics of male and female unemployment rates and highlighted the varying sensitivity of unemployment to economic growth across different EU countries.

Another study [18] examined the relationship between the shadow economy and the unemployment rate in both developed and developing countries using a simultaneous-equation panel data model. The results suggest that in developing countries, there is a unidirectional and negative causality running from the unemployment rate to the shadow economy. However, in developed countries, there is a bidirectional and negative causal relationship between the shadow economy and the unemployment rate. The researchers also found that institutional quality strongly interacts with the relationship between the shadow economy and the unemployment rate. In countries with good institutional quality, the unemployment rate is associated with a weak informal economy, whereas in countries with low institutional quality, it strongly drives the informal economy.

The connection between unemployment and economic growth was analyzed for Jordan during the period from 1991 to 2019 [19]. They employed the auto-regressive distributed lag (ARDL) model for their investigation. The study's findings disclosed a negative relationship between economic growth and unemployment in Jordan. Furthermore, the research pinpointed positive associations between education, the female population, urban population, and unemployment within the country.

A panel data investigation [20] explored the impact of human capital, institutional quality, and the Fourth Industrial Revolution on unemployment rates in 46 Asian countries from 2007 to 2020, utilizing regression techniques on panel data. The study identified nine factors influencing the unemployment rate, including high-tech exports, inflation, population, GDP, government spending, foreign debt, foreign direct investment, and human capital. The research findings provide evidence that increases in inflation, government spending, external debt, and foreign direct investment are associated with higher unemployment rates. Conversely, an increase in population and GDP has the effect of reducing the unemployment rate. Additionally, the study confirms that the Fourth Industrial Revolution contributes to increased unemployment in Asian countries.

2.2. Taxes on income and unemployment

High taxes on corporate profits are normally expected to decrease job offer. Still, that depends on labor productivity but also the investigated period of time. Because, more tax collected at the public budget can support public expenses for education, health, economic and social purposes that all contribute to sustainable economic growth and economic development. So, a sustained economic growth leads to a reduced unemployment rate in the long-term. A study conducted for European economies [21] concluded that corporate taxes have the effect of increasing the cost of capital, subsequently leading to higher levels of equilibrium unemployment. The extent of this impact diminishes with the elasticity of substitution between labor and capital in production, as well as the sensitivity of wages to changes in unemployment. Their analysis further revealed that corporate taxes are particularly burdensome for welfare in European countries that host a significant share of multinational companies and have a relatively open economy. Compared to labor and value-added taxes, corporate taxes have a less detrimental effect on labor market performance. However, they impose more substantial welfare costs in various European countries due to distortions in production efficiency, particularly in open economies like Ireland and the Benelux countries.

Some authors [22] investigated the influence of direct taxation on economic growth by analyzing panel data from all 27 EU countries, spanning the period 2008–2020. The findings revealed that reducing direct taxation has the potential to boost disposable income, stimulate consumption and economic growth, promote investment and job creation, enhance competitiveness, and curtail tax evasion and avoidance, consequently leading to a more efficient tax system. Moreover, they observed that personal income tax was linked to lower economic growth for countries in the limited fiscal efficiency group.

Another study [23] analyzed the impact of the tax structure on economic growth in Croatia using vector error correction model from 2004 to 2019. They found that the impact of different tax structures on economic growth and concludes that direct taxes have a negative effect on growth, while indirect taxes are neutral.

2.3. Social contribution and unemployment

Social contributions collected at the public budget can be directed to finance health and social purposes. This can support increasing the income level and integrating the unemployed people on the labor market again. Previous researchers [24] revealed that investing in human capital through education and health services could significantly decrease unemployment rates in the long term. The research indicated that the variables negatively impacting unemployment are education spending per pupil and health spending. The most effective approach to reducing unemployment is through investments in enhancing the quality of human capital, achieved by funding education and improving health services.

Some studies performed for a large number of countries [25] highlighted the significance of social security contributions as crucial revenue sources for most countries. However, concerns persisted regarding their potential economic consequences, particularly concerning labor supply and employment. Their analysis underscored the critical nature of the link between social security contributions and entitlements. When this connection was clear, employees perceived these contributions more as a price rather than a tax. This nuanced perception resulted in fewer distortions in labor supply, wage costs, and private savings. The study underscored the potential benefits of welfare state reforms that solidified a stronger connection between contributions and individual benefits. Furthermore, they emphasized the need to enhance the visibility of the relationship between social contributions and the associated benefits for citizens.

Researchers [26] revealed that total public social expenditure is negatively associated with poverty and inequality, but it does not show any significant relationship with GDP growth. The study suggests that when social expenditure is utilized to alleviate poverty and inequality, there is no trade-off with economic growth.

Another study elaborated for China [27] examined the impact of the housing provident fund on employment. He discovered that reducing the contribution rate of the housing provident fund could promote employment. However, the negative impacts were pronounced within groups facing strong financing constraints, primarily affecting workers with lower levels of education and male workers. The study also revealed that the fund assisted enterprises engaged in research and development (R&D) in recruiting more employees. Nevertheless, this expansion came at a cost: it lowered the profitability of enterprises, compelling them to scale back their production operations. Consequently, an overall adverse effect on the total number of employees was observed. The paper further found that decreasing the contribution rate of the fund had the potential to significantly augment the workforce within manufacturing enterprises. This trend, in turn, contributed to the overarching objective of employment stabilization.

Previous works [28] investigated the influence of the European Social Fund (ESF) on youth education and employment prospects across various regions in the European Union. Their findings indicate that the ESF has a positive impact on employment rates for youth across all education levels. However, when it comes to education, its impact is more intricate and multifaceted.

According to some studies [29], social spending plays a vital role in enhancing social welfare and mitigating the adverse consequences of unemployment on both individuals and society. The study's results indicate that social spending can have a positive impact on employment levels through the expansion effect, which, in turn, affects the overall demand in the economy. Furthermore, the research highlights that social spending affects chronic unemployment in a distinct and specific manner, and its influence on chronic unemployment is more pronounced in countries with high social spending intensity. This study makes a significant contribution to the existing literature by investigating the link between social spending and chronic unemployment and providing empirical evidence regarding the impact of social spending on chronic unemployment in selected OECD countries.

A study elaborated for EU [30] investigated the effects of pension spending on both poverty reduction and economic growth, utilizing data from 24 European Union Member States covering the period from 2007 to 2018. The findings indicate that pension spending plays a significant role in reducing poverty; however, it does not seem to have a measurable impact on gross domestic product (GDP) growth. The study also considered various control variables, such as the unemployment rate, percentage of population aged between 15 and 64, percentage of population aged 65 or older, and GDP per capita. The results reveal that these control variables have a significant positive influence on the poverty rate.

Economic research [31] highlighted the welfare effects of both existing and counter-factual European unemployment insurance policies. The findings suggested that a harmonized benefit system, which included a one-time payment of approximately three quarters of income upon separation, led to welfare improvement in all euro area countries compared to the then-current status quo. They supported the European policy known as the Temporary Support to Mitigate Unemployment Risks in an Emergency (SURE), which enabled national governments to borrow at low interest rates to cover expenses related to unemployment risks. Furthermore, they advocated for more effective unemployment insurance policies that could enhance employment in the Eurozone.

Previous works [32] found that higher coverage rates of unemployment insurance and minimum income support systems provided a heightened level of income insurance for the unemployed. They suggested that relaxing eligibility criteria for unemployment insurance and minimum income support systems could address income stabilization in the face of macroeconomic shocks and supports employment in the long-run. The researchers investigated the impact of financial development on various components of unemployment in the Middle East and North African countries. Their findings indicated that financial development had a significant negative effect on unemployment across quantiles, although this impact diminished as they transitioned from lower quantiles to higher ones.

2.4. Migration and unemployment

Free movement of factors into integrated EU area can contribute to a sustainable development of the entire area as a whole and the decrease of unemployment. However, that depends on the immigrant skills, productivity and the level of their integration on the host EU markets. Moreover, migrations can reduce unemployment into the host economies, but increase unemployment on the home labor markets. Some authors [33] examined the correlation between cross-border migration and unemployment in Nigeria and Benin Republic for the period 1999 to 2020. The findings indicated a noteworthy negative relationship between cross-border migration and unemployment in Nigeria. However, in the case of Benin Republic, the relationship between cross-border migration and unemployment was negative but not statistically significant. Based on these results, the researcher concluded that cross-border migration played a role in reducing the unemployment rate in Nigeria, whereas it did not seem to have a significant impact on unemployment in Benin Republic.

When the average productivity of immigrants is lower than that of natives, it leads to higher unemployment rates among immigrants [34]. The research highlights that even if the difference in productivity between immigrants and natives is small, the disparity in unemployment rates can still be significant. Furthermore, the study suggests that in certain scenarios, it may be beneficial for the union to prioritize the income of native workers over that of immigrants.

Children of immigrants, particularly those with non-Western backgrounds, tend to achieve higher socio-economic status compared to their parents. However, despite this improvement, they still experience higher unemployment rates compared to their native peers in various countries [35].

2.5. Population and unemployment

Population growth and population aging in the EU area have contributed to high unemployment rates, particularly in Western EU countries, and have led to a significant workforce deficit. The impact of demographic and education changes on unemployment rates in Europe using a panel of European countries for the period from 1975 to 2002 [36]. The researchers concluded that demographic and education changes had a significant impact on unemployment rates in Europe. Changes in the population age structure were positively correlated with the unemployment rate of young workers, while changes in the education structure exhibited a negative effect on the unemployment of the more educated individuals. The study also identified that labor market institutions played a crucial role in enhancing employment opportunities.

Some works [37] conducted an investigation to explore the influence of population growth and unemployment on economic growth in selected West African countries, utilizing data from the World Bank for the period spanning from 1980 to 2019. The study unveiled a noteworthy negative relationship between the unemployment rate and GDP in the West African region. Conversely, population growth was found to have a positive effect on GDP. They emphasized the need to encourage increased investment in the agriculture and manufacturing sectors across the examined region as a means to reduce unemployment. Furthermore, the study recommended implementing appropriate policies to improve the quality of life for the population in the region, thereby fostering economic growth and overall development and higher employment in this region.

3. Model and methodology

This paper investigates the effects of immigrant numbers, income tax and social contributions on unemployment rate. In this regard, data from EU countries between 1991 and 2020 have been utilized.2 Consequently, the subsequent model has been estimated to examine this relationship. This model is based on the analysis performed by previous studies [22] for EU countries, but we have added new variables as social contributions and immigration numbers as per eq. (1):

$${UNM}_{it}={\alpha }_i+{\zeta }_t+{\beta }_1{\text{SOCC}}_{it}{+\beta }_2{\text{MIG}}_{it}+\beta X_{it}+u_{it}$$ (1)

where UNM represents the unemployment rate, Unemployment refers to the share of the labor force that is without work but available for and seeking employment and is published annually by the International Labor Organization (ILO).

MIG is immigrant number, which is the net total of migrants during the period, that is, the number of immigrants minus the number of emigrants, including both citizens and noncitizens.

Also, in this equation, SOCC is the social contributions include social security contributions by employees, employers, and self-employed individuals, and other contributions whose source cannot be determined. They also include actual or imputed contributions to social insurance schemes operated by governments.

Included in the estimation are control variables of income tax, economic growth, population, which are based on the most frequent and prevalent cases in the research's background. All variables accounted for in the estimation Other than the variable immigrant number are expressed in logarithmic form.

The model was estimated using quantile regression method [38]. Traditional regression estimation techniques, such as ordinary least squares (OLS), do not take into account the heterogeneous data distribution, so that in the event of inconsistency (heterogeneity) and aberrant data distribution, the coefficients may be under- or over-estimated [39]. The quantile regression technique permits the estimation of various quantile functions of the conditional distribution, including the median function, which represents a special state. Each quantile regression specifies a unique point (tail or center) of the conditional distribution. Placing distinct quantile regressions side-by-side provides a more comprehensive distribution of the original conditional distribution. In actuality, this model is used to overcome the limitations of conventional regression models, and it has been extensively adopted due to its advantages over conventional regression methods. The ability to provide real and accurate findings in the presence of outlier data [40], the flexibility of coefficients throughout the distribution [41], freedom from problems related to sample selection bias [42], and the ability to explain the effect of independent variables (regressors) on the dependent variable in different quantiles [43] are among its most significant advantages. Several economic studies have used the quantile regression method to investigate the relationships between variables [32,44,45]. Estimating in different quantiles is important in economics because the relationship between variables is often non-linear. It changes according to different quantiles, so the policy measures can be adapted to that.

Different quantile functions are estimated using a conditional distribution in this method. It distinguishes itself from the ordinary least squares method in that the latter examines the average behavior of the relationship between the dependent and independent variables, whereas the quantile regression examines this relationship at different levels and different quartiles. In equation (2), the estimation of quantile regression minimizes the flowing objective function described below:

$$Q\left(\beta q\right)={\sum }_{i:yi>x^{\prime }\beta }^{n}q|yi-x^{\prime }\beta q+{\sum }_{i:x^{\prime }\beta }^n\left|yi-x^{\prime }\beta q\right|$$ (2)

Using the quantile panel regression method, this research investigates the impact of immigrants number and social contributions on unemployment rate in EU countries. The conditional quantile function presented in this study is presented as eq (3).

$${UNM}_{it}={\alpha }_i+{\zeta }_t+{\beta }_1{socc}_{it}+{\beta }_2{MIG}_{it}+{\beta }_3{TAXIN}_{it}+{\beta }_4{GDPPC}_{it}+{\beta }_5{POP}_{it}+u_{it}$$ (3)

To investigate the effect of the effective factors on the various quantiles of the dependent variable, equation (4) is written as follows:

$$Q_{{UNM}_{it}}\left(\tau |{\alpha }_i,{\zeta }_t,X_{it}\right)={\alpha }_i+{\zeta }_t+{\beta }_{1\tau }{socc}_{it}+{\beta }_{2\tau }{MIG}_{it}+{\beta }_{3\tau }{TAXIN}_{it}+{\beta }_{4\tau }{GDPPC}_{it}+{\beta }_{5\tau }{POP}_{it}$$ (4)

where Q_t represents the regression parameter of the $\tau$ ^th quantile in the dependent variable, whereas ${\beta }_{1\tau }$, ${\beta }_{2\tau }$, ${\beta }_{3\tau }$, ${\beta }_{4\tau }$ and ${\beta }_{5\tau }$ represent the regression parameters of the $\tau$ ^th quantile in the explanatory variables.

3.1. Panel cointegration test

Once the stationarity of the studied time series is confirmed, it becomes valuable to determine the presence of a long-term co-accumulation relationship among the variables of interest. In this research, three commonly used tests, namely Pedroni [46] and Kao [47], were employed, but also Westerlund which provides robust results in case of cross-sectional dependence. Pedroni [46] introduced two types of cointegration tests: panel test statistics and group test statistics. The panel tests are grounded in the intra-dimensional approach, encompassing four statistics: panel V-statistics, panel Rho-statistics, panel PP-statistics, and panel ADF-statistics. Group tests are rooted in the interdimensional approach, comprising three statistics: group rho statistic, group PP statistic, and group ADF statistic. Additionally, the Kao panel cointegration test follows the same estimation method as the Pedroni test, but incorporates cross-sectional intercepts and homogeneity coefficients in the regression [48]. The null hypothesis of these tests presumes the absence of a cointegration relationship among the variables of interest, as opposed to the alternative hypothesis that posits the presence of a cointegration relationship.

3.2. Cross-section dependence tests

The cross-section dependence (CD) test proposed by Pesaran [49] tests the null hypothesis of zero dependence across the panel members and is applicable to a variety of panel data models such as stationary and unit root dynamic heterogeneous panels with structural breaks, with small T and large N [49]. The CD test is based upon an average of all pair-wise correlations of the ordinary least squares (OLS) residuals from the individual regressions in the panel data model (eq. (5)):

$$y_{it}={\alpha }_i+{\beta }_i{x}_{it}+u_{it}$$ (5)

where i = 1, …, N represents the cross-section member, t = 1, …, T refers to the time period, and x_it is a (k × 1) vector of observed regressors. The intercepts, αi, and the slope coefficients, β_i, are allowed to vary across the panel members. The CD test statistic is defined as per eq. (6):

$$CD=\sqrt{\frac{2T}{N\left(N-1\right)}}\left(\sum _{i=1}^{N-1}\sum _{j=i+1}^N{\hat{\rho }}_{ij}\right)\to N\left(0,1\right)$$ (6)

Where ${\hat{\rho }}_{ij}$ is the sample estimate of the pair-wise correlation of the OLS residuals, ${\hat{u}}_{it}$, associated with Equation (7)

$${\hat{\rho }}_{ij}={\hat{\rho }}_{ji}=\frac{{\sum }_{t=1}^T{\hat{u}}_{it}{\hat{u}}_{jt}}{{\left({\sum }_{t=1}^T{u}_{it}^2\right)}^{1/2}{\left({\sum }_{t=1}^T{u}_{jt}^2\right)}^{1/2}}$$ (7)

4. Results

In Table 1 we have presented the variables we have used into the estimation model, their unit of measure and data source. In Table 2 we can see that unemployment and immigration flows are nor normally distributed and present outliners.

Table 1.

Variables and data source.

Variable	Description	Source
UNM	Unemployment, total (% of total labor force) (modeled ILO estimate)	World bank https://databank.worldbank.org/source/world-development-indicators
SOCC	Social contributions (% of revenue)	World bank https://databank.worldbank.org/source/world-development-indicators
TAXIN	Taxes on income, profits and capital gains (% of total taxes)	World bank https://databank.worldbank.org/source/world-development-indicators
GDPPC	GDP per capita (constant 2015 US$)	World bank https://databank.worldbank.org/source/world-development-indicators
MIG	Net migration flows (million)	World bank https://databank.worldbank.org/source/world-development-indicators
POP	(million)	World bank https://databank.worldbank.org/source/world-development-indicators

Source: compiled by authors

Table 2.

Descriptive statistics of variables.

	UNM	SOCC	TAXIN	GDPPC	MIG	POP
Mean	8.3632	31.3603	38.4355	10.0353	22070.7453	15.7190
Maximum	27.4699	68.5128	60.1225	11.6299	859739	18.2362
Minimum	1.4800	1.9598	9.54273	8.30825	−138327	12.8044
Skewness	1.2359	−0.1490	−0.4090	−0.1607	4.4095	−0.1716
Kurtosis	5.0337	3.3881	2.5702	2.4881	33.1172	2.5943

Source: calculated by authors

The Jarque-Bera test is used to detect the normality of the data. The Jarque-Bera normality test can be used to analyze the skewness and skewness of the coefficients. According to the results of the normality test of the residuals, the results of Table 3 show that they are not normally distributed.

Table 3.

Test of normality of residuals.

Jarque-Bera statistic	298.761
probability	0.000

We have applied Pesaran CD test which validated the CSD among the panel. Therefore, it is necessary to employ second generation estimation techniques, so that the results can't be biased (Table 4).

Table 4.

Cross-sectional dependency test results.

	CD	p-value
UNM	19.306	0.000
SOCC	2.513	0.012
TAXIN	2.415	0.016
GDPPC	81.335	0.000
MIG	8.502	0.000
POP	9.648	0.000

Source: calculated by authors

According to the Pesaran Unit Root test [49] presented in Table 5, all variables are integrated I(1), except immigration flows which is integrated I(0). This mix order of integration requires an ARDL method of estimation as robustness test, after performing quantile regression.

Table 5.

Pesaran Unit root test result.

Variable	level	p-value	First difference	p-value
UNM	−1.602	0.055	−5.666	0.000
SOCC	−0.728	0.233	−6.534	0.000
TAXIN	1.443	0.925	−4.466	0.000
GDPPC	−0.748	0.227	−3.117	0.001
MIG	−1.800	0.036	−3.982	0.000
POP	2.918	0.998	−3.852	0.000

Source: Author's Calculation

We have applied Westerlund, Pedroni and Kao co-integration test to check if there is a long-run relationship among variables and the H0 hypothesis is validated at 1 % and 5 % significance level respectively (Table 6).

Table 6.

Panel-data cointegration tests.

	Statistic	p-value
Kao	−2.2545**	0.0121
Pedroni	3.7383***	0.0001
Westerlund	−1.8531**	0.0319

Note: ***,** and* means 1 %, 5 % and 10 % significance level.

Source: Author's Calculation

In Table 7 we have presented the results of quantile regression estimations. Income tax and social contributions are not statistically significant at lower quantiles. The impact of income tax can be observed much sooner, while the impact of social contributions on unemployment rate can be noticed only in quantile 80th and 90th. The impact of income tax on unemployment is negative, just like the impact of the social contributions on unemployment is negative. That means that at higher levels of income taxation and social contributions, the unemployment rate decreases. Population growth, GDP per capita and immigration flows are statistically significant in all quantiles. That means that no matter of their level, they impact on unemployment rate in EU countries. Population and GDP per capita display strong and significant impact on unemployment in EU countries. Population is positively associated to unemployment rate which means that a population growth determines the rise of unemployment rate, while GDP per capita is negatively associated to the unemployment rate. If GDP per capita increases, the unemployment rate will decrease because the boom of the economic activity generates new employment options. Immigration flows impact is weak and negatively associated to unemployment rate. That means that the free movement of labor force inside EU can determine the decrease of the unemployment rate of the EU countries, but this impact is not definitory for the development of the unemployment rate among EU countries (see Table 6).

Table 7.

Quantile regression results.

	Quantile	0.1	0.2	0.3	0.4	0.5	0.6	0.7	0.8	0.9
TAXIN	Coefficient	−0.010	−0.040***	−0.040***	−0.047***	−0.046***	−0.046***	−0.061***	−0.126***	−0.206***
	Std. Error	0.021208	0.013486	0.011625	0.010277	0.010643	0.012835	0.017248	0.020086	0.024129
SOCC	Coefficient	−0.014	−0.006	0.002	−0.003	−0.007	−0.011	−0.025	−0.089***	−0.201***
	Std. Error	0.012387	0.007997	0.008304	0.007659	0.009738	0.015642	0.018013	0.023566	0.032433
POP	Coefficient	0.545***	0.625***	0.703***	0.807***	0.794***	0.948***	1.242***	1.603***	2.060***
	Std. Error	0.064238	0.064428	0.092675	0.076771	0.075306	0.076484	0.07691	0.087961	0.131696
MIG	Coefficient	−3.21E-06**	−4.00E-06***	−5.30E-06**	−6.95E-06***	−6.96E-06***	−7.88E-06***	−1.13E-05***	−1.23E-05***	−1.30E-05**
	Std. Error	1.39E-06	1.18E-06	2.15E-06	2.49E-06	2.47E-06	2.50E-06	2.47E-06	2.66E-06	5.15E-06
GDPPC	Coefficient	−0.333**	−0.264**	−0.327**	−0.369***	−0.268**	−0.412***	−0.646***	−0.616***	−0.442**
	Std. Error	0.13789	0.114793	0.146412	0.11933	0.123876	0.103021	0.116698	0.146303	0.214489

Note: ***,** and* means 1 %, 5 % and 10 % significance level.

Source: Author's Calculation

To ensure robustness, we have applied PMG-ARDL method as a result of co-integration and first order integration of the variables we have use into the model. ARDL models are least squares regression models in which dependent and independent variables are used as explanatory variables. In case of using panel data models with individual effects, using the standard ARDL estimation method will not give correct results due to the existence of correlation between explanatory variables and error components, and the estimation will be biased and inconsistent. This strain does not disappear either by increasing the time dimension or by increasing the sections. To correct this problem in panels with a small-time dimension and a relatively large cross-sectional dimension, the GMM method was developed by Ref. [50].

However, this method cannot be used in panels with a large time dimension. In this case, the PMG estimator that was presented by Pesaran et al. [51] can be used. This model takes the cointegration relationship from the simple ARDL model and then adjusts this model between different cross-sectional units.

In this study, to check the robustness of the model, the pooled mean group (PMG) estimation method proposed by Pesaran et al. [51] was used. PMG estimation is a useful intermediate method between two limit methods. These two limit methods that are often used to analyze dynamic panel models are dynamic fixed effects estimator (DFE) and group mean estimator (MG). The DFE estimator imposes the strong homogeneity assumption that all short-term and long-term coefficients and error variances are the same across countries. On the other hand, the MG model estimates the short-term and long-term coefficients for each country separately and examines the distribution of the countries' estimates - usually their average. In the PMG method - the middle state of these two limit states - the short-term coefficients and variance of errors are different between countries, but the long-term coefficients are similar between countries. As a result, PMG estimators provide better estimates than the MG method. At the same time, they pay attention to the heterogeneity between countries in short-term dynamics which is important in panel estimations and they are suitable when variables data series are not normally distributed [52].

In Table 8 we have displayed the results of estimations of PMG-ARDL used as robustness test. These results validate the quantile regression results for population, immigration, GDP per capita and income tax. The impact of these specified variables on unemployment rate is strong and significant and the sign of those variables is the same as per estimations of quantile regression. The impact of immigration flows is still very weak and negative. Indeed, the immigrants can support the decrease of unemployment rate in the host countries, but in their home countries, the unemployment rate increases, so the overall impact on unemployment rate in EU is weak. Also, the impact of immigration on unemployment depends on the productivity of immigrants. This is especially true because many immigrants from the Eastern European countries work in the Western European countries, while in the Eastern countries, many Asian immigrants have started to be employed lately for jobs that were not attractive for the local inhabitants. This phenomenon is also true for the South and North of Europe. However, the impact of immigration flows on unemployment rate is very weak. The main difference against quantile regression estimations can be observed for the social contributions. Their impact was proved to be significant and negative in upper quantiles, while in PMG-ARDL estimations, its impact is weak but positive. So, if we don't analyze in each quantile, the overall impact of social contributions on unemployment is positive, but weak. That means that only for high levels of social contributions, they can reduce unemployment, but overall, a certain increase of social contribution level determines the increase of unemployment, because the employers have a big burden in terms of social contributions to the public budget and they will reduce jobs offer. High tax on profits determines firms to create jobs, especially when the labor productivity is high, so that they can diminish their gross profits. This explains the negative relation between tax on profits and unemployment rate.

Table 8.

PMG-ARDL estimations results.

Variable	Coefficient	t-statistic
Pop	2.407789***	5.814038
Mig	−0.0000509***	−13.65639
Gdppc	−12.25328***	−17.80673
Taxin	−0.131638***	−5.239422
socc	0.082204***	2.972989
c	98.46709***	12.90175
Ecm(-1)	−0.135509***	−3.130797

Note: ***,** and* means 1 %, 5 % and 10 % significance level.

Source: Author's Calculation

5. Discussion of results

Our results are validated by the findings of previous studies. Many researchers investigated European or Asian countries and demonstrated a negative association between economic growth and unemployment, because a boost of the economic activity will generate new jobs so that the unemployment will be reduced [14,15,17,19,20].

The negative association between tax on profits and unemployment rate was also found by some previous studies [21]. These have showed for EU countries that corporate taxation is detrimental to the economic activity and thus increases unemployment rate, but this impact depends on the substitution between labor and capital in production and on the elasticity of wages to the development of unemployment rate. They also concluded that the impact of corporate tax on economic activity is higher for economies that are highly open, with many multinationals working on those markets. Wages tend to be aligned inside EU area so the elasticity of wages to unemployment rate has strongly decreased after the settlement of an enlarged EU area. Multinationals are interested in corporate tax on profits when locating on a foreign market, so this tax is very important for those large corporations that create jobs on emerging markets. Tax on personal income is only important for immigration process, and this immigration variable was demonstrated to have a very weak impact on unemployment rate in EU area.

Our results for the association between social contributions and unemployment rate are mixed. We have found a negative association for higher quantiles but an overall positive association with unemployment rate. However, those results are supported by the findings of previous studies [24,27] that demonstrated that higher social contributions determine a decrease of unemployment because they can support increased expenses for health or education, or expenses for reducing poverty and social inequality. In China, it was found a negative association only for some firms, involved in R&D activities, but an overall increase of unemployment as a result of higher social contributions that increase the total cost of enterprises [27]. Some authors [28,31] demonstrated a negative association for EU countries. Same negative association was proved for OECD countries [29] or for Middle East and North-African countries [32].

Our findings about the negative association between immigration and unemployment for all quantiles was also validated by previous studies [33] that demonstrated same negative association between cross-border migration and unemployment rate in Nigeria. It was concluded that this relation can be explained by the differences among the productivity of natives against immigrants, but also by the fact that immigrants always ask for much lower salaries against natives [34,35]. Some authors [53,54] demonstrated positive outcomes of immigration on economic growth and employment in EU area, meaning that large immigration flows can determine the decrease of unemployment rate in EU. However, the impact of immigration across EU countries on unemployment is weak because the poorer EU countries face a large labour deficit because of large labour force outflows.

Growing population obviously determines the increase of unemployment all over the world. Because it requires some adaption of the economic structures and of educational systems to this strong growth of population so it can meet the labor market needs. This positive relation between population and unemployment rate was demonstrated by other authors as well [37]. These demonstrated a positive association between population growth and reducing economic activities, while increasing unemployment for West African countries. Some studies [36] demonstrated a strong and significant association between population growth and unemployment rate in EU countries and proved that the developments of age structure of the population determined the increase on unemployment especially for young people in EU countries.

Summarizing our results, a robust economic growth will support creating jobs and also higher tax on corporate profits can determine organizations to be interested in hiring people especially if the labor productivity is high. However, taxation on profits and supporting economic activity are two opposite targets of the economic policy. Considering the fact that the impact of the economic growth on unemployment is stronger than the impact of tax on corporate profits, it is obvious that the main goal of the authorities of the EU countries should be achieving a strong economic growth. Also, they need to use social contributions revenues collected at the state budget to finance social purposes to reduce social inequality and to enhance training and education of unemployed people so they could be re-inserted quickly on the labor market. The weak impact of immigration on unemployment rate shows that in the long-run it can't solve the unemployment issue in the EU area, despite free movement of labor force. Overall, the income collected at the state budget can be directed to economic purposes, to support business and insure a robust economic growth for decreasing unemployment rate. Also, they can be used for designing more active policies on the labor markets for supporting the insertion of unemployed people on the labor markets again, by organizing training and professional reconversion programs to people without jobs. Of course, the target of reducing unemployment significantly and in the long-run can't be achieved without the increase of public debt as it could been noticed during the last pandemic. Immigration inside EU borders can't prove to be a solution, because large discrepancies in the economic growth rates among EU states and fluctuations of the growth rates make immigrants very vulnerable to the economic internal conditions of each EU country. This was validated during the last pandemic when many immigrants returned to their home countries because they have lost their jobs abroad because the economic activity stopped or was suddenly reduced all over the world and because prices exploded. Fiscal conditions became irrelevant in that macroeconomic context. Moreover, Asian immigrants can represent a better and less costly substitute for European immigrants in some domains. So, the focus shouldn't be only on cutting the tax on private sector and public indebtness, but rather on how can collected incomes at the public budget can be used to support economic activity and to support the implementation of active labor policies. Also, public production should be stimulated for decreasing unemployment. Fiscal competition among EU member states and immigration flows inside EU borders don't represent solutions for decreasing unemployment rate in EU area in the long-run.

6. Conclusions and policy recommendations

In this study we have aim to investigate the relationship between immigrants, social contribution paid to the public budget and income tax on profits on the unemployment rate in EU countries during 1991–2020. Additionally, we have employed GDP per capita and population growth as control variables. For this purpose, we have checked cross-sectional dependence among the EU panel based on Pedroni CD test. The CSD hypothesis was validated by the results of the test. We have applied Pesaran unit root test and found that all variables are integrated I(1) except migration which is I(0). We also examined the co-integration of the variables in the long-run by applying Pedroni and Kao tests and the co-integration was validated. Subsequently, we have explored the relation between exogenous variables and unemployment rate using a panel quantile regression. For ensuring robustness of our results, we have applied PMG-ARDL method for a robustness check.

The results demonstrate that tax on income (on profits) exerts a negative impact on unemployment rate in all quantiles, except the 10th quantile, while social contributions are statistically significant only for higher quantiles 80th and 90th quantiles respectively. In these quantiles, the impact of social contributions is negative. The number of immigrants is significant in all quantiles, but its impact is very weak and negative. Both population and GDP per capita are significant and display a strong impact in all quantiles, but the impact of population on unemployment rate is positive, while the impact of GDP per capita is negative. PMG-ARDL estimations validate the results of quantile regression, except for social contributions. This variable displays a positive impact, but weak. The overall positive impact of social contributions on unemployment rate can be explained because the negative impact was demonstrated only for higher quantiles cause this way the public expenses for social purposes can increase and support the decrease of unemployment in the long-run. Its overall positive impact can be explained by the rising cost of enterprises that pay higher social contributions to the public budget and that affect the overall economic activity and thus, the unemployment rate.

Based on the achieved results, we can emphasize that tax on corporate profits and economic growth can support reducing unemployment rate among EU area. Furthermore, immigrants should be integrated on the host labor markets according to their abilities, their studies should be certified accordingly and some training and insertion programs should be implemented for achieving better labor skills for them. Only that way the immigration can support reducing the overall unemployment into EU enlarged area on sustainable basis. Stimulating economic growth and setting higher tax on profits are two targets difficult to achieve in the same time. This is because corporations are very sensitive to profit taxation. Therefore, labor productivity is a factor that should be taken into consideration in this discussion. However, higher tax at the public budget can be used for education, health, social and economic purposes so that they can contribute to achieving economic growth in the long run and decreasing unemployment rate.

A limitation of this study is represented by the period of time investigated into this study, due to the availability of data. Analysis can be extended on sub-groups of immigrants, those coming from European countries or those coming from other geographical areas, because they are very different in terms of insertion on EU labor market, their motivation or their productivity. Also, some other variables should be included into the estimation model, such as labor productivity and tax on personal income. Institutional variables could also provide valuable insights for an analysis of unemployment rates. Economic freedom, governance efficiency and labor union density can be important institutional variables for labor market and for the development of unemployment rate.

Ethical approval

Not applicable.

Consent to participate

Not applicable.

Consent for publication

All authors approved the manuscript to be submitted for publication.

Funding

This research received no funding.

Data availability statement

All data are publicly available and authors provided their source into the manuscript. Data are collected from World Bank database and links are provided in Table 1.

CRediT authorship contribution statement

Ali Moridian: Writing – original draft, Software, Methodology, Formal analysis, Conceptualization. Magdalena Radulescu: Writing – review & editing, Validation, Supervision, Project administration, Conceptualization. Parveen Kumar: Writing – original draft, Investigation. Maria Tatiana Radu: Writing – review & editing, Investigation. Jaradat Mohammad: Writing – review & editing, Data curation.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.