Forecasting Sustainable Development Indicators in Romania: A Study in the European Context

: Sustainable development is a very important and highly debated concept worldwide, with almost all states around the globe recognizing the importance of supporting a balancing of economic, social and ecological plans to create a more prosperous and equitable world. The purpose of the current study is to identify the current context of the progress achieved by the EU member states regarding the 17 sustainable development objectives. Following the implemented research, it was found that Romania is one of the countries with the lowest performance, so an attempt was made to make a short-term forecast in the case of this country in order to determine its future course. In this sense, the Holt–Winters trend methods were used, which are based on historical data to predict future values. According to the analysis carried out, the predicted values are mainly encouraging, but it is necessary for the competent authorities to become more involved because, in many aspects, Romania is at the bottom of the European Union (EU) ranking regarding the analyzed indicators.


Introduction
Sustainable development is a concept of global interest aimed at balancing economics and social and ecological issues.In this sense, in 2015, the 193 member states of the UN signed the 2030 Agenda, among their wishes being the need to eradicate poverty, protect the planet, limit inequalities and provide a safer environment for humanity [1].
The primary objective of this paper is to examine the European context in relation to various targets outlined in the 2030 Agenda.In this sense, it was possible to determine Romania's position regarding the progress made in achieving these objectives in relation to the other member states, noting that, in many cases, it is at the bottom of the ranking, raising the question of whether or not it will be able to reach the proposed targets by 2030.Thus, an attempt was made to forecast a possible trajectory to observe the targets that are not yet on the desired path.Subsequently, an analysis of historical data from Romania is conducted to forecast the three-year progression of the indicators under study.These indicators include factors such as the population at risk of poverty or social exclusion, individuals who leave education and training prematurely, the Corruption Perceptions Index and the representation of women in senior management positions.The Crystal Ball Oracle extension was used to make the forecasts, the models used for this study being Holt-Winters trend methods.To ensure the soundness of the predictions, those models were chosen for which the RMSE had the minimum value, and to test the accuracy of the forecasts created, Theil's U and Durbin-Watson tests were used.
The 17 Sustainable Development Goals emerged as part of the 2030 Agenda to create some targets to be achieved for building a world that enjoys peace and prosperity.Considered to be in a global partnership, they were drawn up to continue what could not be achieved through the Millennium Development Goals but also to improve the world in which we live.The world leaders have realized that eradicating poverty can only be achieved by improving medical and educational conditions, reducing gender and social inequities, promoting equal opportunities and protecting the environment that provides us with food [2].
This concept of sustainable development has received significant attention from researchers.In a bibliometric analysis conducted by Vatananan-Thesenvitz et al. [3], it was found that, between 1985 and 2019, a total of 16,040 sustainability-related documents were published in the SCOPUS database.Furthermore, it is worth noting that the field of innovation alone contributed significantly to this body of literature [3].
Starting with innovation and research, it is noteworthy that their impact on economic growth is positive, particularly for countries that have recently become part of the European Union (EU).According to a study conducted by Pelinescu et al. [4], it is shown that, on average, at the EU level, an increase in research and development expenditures (per capita) coupled with a rise in the number of granted patents contribute to enhancing economic growth.Furthermore, a positive trend has been highlighted in the scientific literature for renewable energy consumption, both at the EU and Romanian levels, within the last period of time [5].Oprea et al. [6] introduced a method for reducing energy consumption, noting that shifting usage to off-peak hours can cut peak consumption by over 23%.Other studies expanded on this by identifying suspicious consumers [7] and detecting anomalies [8].Solutions for managing electricity use have been proposed for individuals [9], commercial consumers [10] and entire markets [11][12][13][14].
In this paper, values for the targets of the sustainable development objectives will be forecasted; thus, the progress made in the world with regard to these objectives must be observed so that the analysis will then focus on the local level, initially starting from the entire EU and subsequently describing the case of Romania.Moyer and Hedden [15] wrote an article in which the global forecast for six of the seventeen sustainable development goals is described, focusing on people's access to education, food and basic utilities.For Europe, the targets will have achievement thresholds greater than or equal to 80%.
Another relevant example in this context is the article authored by Paraschiv et al. [16].The study utilizes the Holt-Winters method to forecast future values specifically in the textile industry.The authors describe the fact that this area of production causes more than 20% of the water pollution in Eastern European countries.Thus, since it is desired that the degree of pollution decreases and the industry focuses on a more sustainable way of production, the authors began by describing the time series for water pollution caused by the textile industry and for total emissions of organic water pollutants for Romania and Poland (the countries with the highest values in Eastern Europe at the time).The authors noticed that, in the case of Romania, water pollution caused by the textile industry had an upward trend from 1998 to 2004, while in Poland, the trend was one descent for both indicators analyzed in the period under observation.According to the forecast performed in the study for the period 2007-2017, a downward trend is expected for both countries, moving toward complete elimination of water pollution caused by the production process of this industry [16].
The Holt-Winters method has been used individually or together with other methods in fields strictly linked to sustainable development or related to it, such as health [17,18], energy [19,20], pollution [21] and agriculture [22], but not limited to these [23,24].
The interest given to the identification of the route to be taken towards the achievement of these objectives was concentrated both at the global [25] and EU levels [26], locally-the member states [27,28]-as well as outside the EU area, the objectives being analyzed together [15], then at the objective level [29], but also only certain targets that compose various objectives [30,31].
Another effective approach to reducing pollution is carbon neutralization.A study conducted by Wang et al. [32] analyzed European countries during the period of 2001-2019.The research findings indicate that, when the innovation level reaches a threshold of 10.5, a 1% increase in innovation results in a significant decrease of 1.896% in carbon emissions.Similarly, for carbon taxes set at a threshold level of 12.80, it is observed that a 1% increase in these taxes leads to a 1.697% reduction in emissions.Furthermore, if this threshold is surpassed, the same 1% increase in carbon taxes is associated with an approximately 3% decrease in carbon emissions.
Healthcare is another significant aspect within the social plan, and the well-being of citizens is a key concern.Diabetes poses both physical and financial challenges for individuals affected by this condition [33].In their study, the authors conducted a forecast of economic implications, specifically the costs associated with the care of patients with diabetes, for the year 2030.The forecast was based on three distinct scenarios that provided information on the potential financial impact of diabetes on the health system.The described cases assume that mortality and prevalence increase primarily with urbanization and population aging, this being the baseline scenario.The second case refers to previous trends, and the last describes the case in which the objectives of the 2030 Agenda and World Health Organization Global Action Plan for the Prevention and Control of Noncommunicable Diseases (2013-2020) are achieved.In all three cases, the financial burden is very high, and the differences between them are not so significant.Thus, according to the authors, a series of policies should be implemented and promoted to prepare the medical and social systems to mitigate the effects of diabetes.
The predictability of certain aspects, such as the economic impact of diabetes, contrasts with the unpredictable nature of events like the COVID-19 pandemic.People's perceptions of their health can significantly influence their well-being, and during an unforeseen pandemic, individuals are profoundly affected.It is noteworthy that informal forecasting by public health practitioners was not regarded as significant in the context of the pandemic.According to a study carried out in England by Davies and Ferris [34], the majority of practitioners surveyed were able to correctly predict some of the public measures taken to stop the increase in the number of COVID-19 cases, their opinion on what can happen at the local level being very important because each local practitioner knows how many infected people he or she had to manage, the measures taken by them being very important at the local level.
Based on the selected paper, it becomes noticeable that numerous authors seek to explore the potential future progress that can be achieved in relation to sustainable development indicators.Whether the focus is on addressing the pollution caused by specific industries [16], finding effective methods to neutralize carbon emissions [32] or emphasizing the significance of population health and well-being [33], all of these studies share a common primary goal: to raise awareness among the population regarding the potential risks that may arise if the authorities fail to implement the necessary measures to achieve the objectives outlined within the 2030 Agenda.
While previous studies share the common goal of forecasting the trajectories of indicators related to sustainable development, they often focus on specific components, such as health or industry, without providing a comprehensive overview.Additionally, while sustainable development is a global concept, it is essential to examine local progress.Some previous studies have explored the global or EU perspective without delving into individual country's situations.Thus, our current research aims to address these gaps by offering a detailed analysis of all Sustainable Development Goals (SDGs) within the context of one country: Romania.This approach is motivated by Romania's predominantly poor performance in achieving the SDGs within the EU context.By focusing on Romania, we seek to provide a nuanced understanding of the challenges and opportunities for sustainable development within a specific national context.

Materials and Methods
In order to assess the analysis aimed in this paper, the double exponential smoothing or Holt's linear trend method [35] has been used.This method involves an equation in which both the level and the trend are included [36]: where S t is the local level and X t is the current value; where T t is the trend and α and γ are the smoothing parameters (having values from 0 to 1); where X t (m) is the predicted value at time point m + t.
For non-seasonal damped trend, a new parameter is added to the above equations [36]: In this case, it is noted that the trend adjusts with this new value ϕ.If ϕ is 0, then the equation is single exponential smoothing.For the case where ϕ is equal to 1, the model becomes double exponential smoothing.If it is greater than 1, the trend is exponential, and for values of ϕ between 0 and 1, the trend is damped equations [36].
Theil's U test is used to determine whether or not the model is better than guessing.The formula for it is replicated below [37]: When the value is lower than 1, the prediction is better than guessing.If the value is equal to 1, the prediction is as good as guessing.If the value is higher than 1, then the prediction is worse than guessing.
The Durbin-Watson test is used to determine the autocorrelation of the errors at lag one.The values for this test are in the range [1,4], and their meaning is as follows [38]: if the value is less than 1, then the errors are positively correlated; when the value is 2, there is no autocorrelation; if the value is greater than 3, then the errors are negatively correlated.The test formula is as follows [39]: where e t is the difference between ŷt and y t and n is the number of periods.Since the number of available years is limited, the Crystal Ball extension was used for the current research, which analyzes the data and, based on the identified patterns, indicates a series of methods that can be used to explain the historical data [40].By means of this extension, those models are calculated for which the error between the predicted and actual data is minimal (the user can choose between root mean square error, mean absolute deviation and mean absolute percentage error).Other methods for checking the accuracy of the provided models are the ones already presented, respectively Theil's U and Durbin-Watson.These two tests are used to check the goodness of the chosen model to ensure prediction accuracy and independence from past errors.

Data and Results
In this section, the European context for 2021 will be presented on the indicators of the 17 Sustainable Development Goals, and a three-year forecast for Romania will be developed using the Holt-Winters trend method.

Descriptive Statistics-The Member States of the EU
For the descriptive analysis, one target for each objective of sustainable development was chosen as an indicator.The data for EU27 countries were collected from Eurostat or the World Bank for the year 2021 with the exception of objectives 6 and 15 for which the last data available were from 2020.For the PM2.5 air pollution, data were collected from IQair.
As stated in the section dedicated to specialized literature, the 17 objectives are grouped into three major plans, economic, social and ecological, all of which are integrated through objective 17, namely Partnership for the Goals.The ecological plan includes objective 6 (Clean water and sanitation), objective 13 (Climate action) and objectives 14 and 15, which refer to Life below water and on land.The social pillar encompasses approximately half of the objectives; here are objectives from 1 to 5 (No poverty, Zero Hunger, Good health and well-being, Quality educations and Gender equality), objective 7 (Affordable and clean energy), objective 11 (Sustainable cities and communities) and objective 16 (Peace, justice and strong institutions).And at the economic level, the other remaining objectives can be found, namely 8 (Decent work and economic growth), 9 (Industry, innovation and infrastructure), 10 (Reduced inequalities) and 12 (Responsible consumption and production) [41].
For the percentage of people at risk of poverty or social exclusion, it can be observed, in Table 1, that the coefficient of variation is less than 30%, a sign that the data are homogeneous; also, the average is 20.62%, the values deviating, on average, from this value by 5.51%.In the case of Government support to agricultural research and development, it is noted that, between the minimum and maximum values, there is a considerable difference from 0.10 to 23.80, the average being 6.57 euros per inhabitant, the coefficient of variation exceeding 80%, being a consistent value with the data discrepancy.Regarding positions held by women in senior management positions (board members), there is also a big difference between the two extreme values (minimum and maximum), the average at the level of the EU being 26.45%.For the percentage of the population unable to keep their home adequately warm, the coefficient of variation has a value of 93.05%; thus, it can be stated that the data are heterogeneous, the average (7.37%) being, in this case, much closer to the minimum value (1.30%) than to the maximum (23.70%) in fact, which is encouraging.Also, the coefficient of asymmetry is positive, a sign that small values prevail; therefore, several countries have a small percentage of the population that cannot heat properly.The OB1 defines people at risk of poverty or social exclusion, and in Figure 1, it can be observed that Romania is the country with the highest percentage of these people, followed by Bulgaria; at the opposite pole are the Czech Republic, Slovenia and Finland.Regarding OB 7, Bulgaria and Lithuania are the countries with the largest number of people unable to keep their home adequately warm, their percentage exceeding 20% for the year under study.However, approximately half of the countries have a percentage of less than 5% for this indicator.The OB1 defines people at risk of poverty or social exclusion, and in Figure 1, it can be observed that Romania is the country with the highest percentage of these people, followed by Bulgaria; at the opposite pole are the Czech Republic, Slovenia and Finland.Regarding OB 7, Bulgaria and Lithuania are the countries with the largest number of people unable to keep their home adequately warm, their percentage exceeding 20% for the year under study.However, approximately half of the countries have a percentage of less than 5% for this indicator.Figure 3a describes the perception of the citizens of each European country regarding health (OB3) and corruption (OB16).Thus, according to the opinion of the population, the most corrupt countries are Bulgaria, Romania and Hungary, and the least corrupt are Finland, Denmark and Sweden.Regarding health perception, among the countries with the lowest percentages of people who consider their health to be good are Lithuania, Latvia and Portugal, and those with high percentages are Ireland, Greece and Cyprus.
OB4 designates early leavers from education and training, and according to Figure 3b, the highest percentage of these people is in Romania, followed by Spain, Bulgaria and Italy, while the countries with a small percentage of the population for this indicator (under 4%) are Croatia, Greece, Ireland and Slovenia.As for Young people neither in employment nor in education and training, Italy, Romania and Bulgaria also have the highest percentages; for the first two countries, the values exceed 20%.The countries with low percentages of this indicator are the Netherlands, Sweden and Slovenia.From Figure 4a, it can be observed that the country with the most positions held by women in senior management positions (board members) is France, where the percentage exceeds 45%, followed by Italy and the Nordic countries.The countries where this percentage does not exceed 10% are Cyprus, Hungary and Spain.For the income share of the bottom 40% of the population, it is observed that Italy and Romania have the highest Figure 3a describes the perception of the citizens of each European country regarding health (OB3) and corruption (OB16).Thus, according to the opinion of the population, the most corrupt countries are Bulgaria, Romania and Hungary, and the least corrupt are Finland, Denmark and Sweden.Regarding health perception, among the countries with the lowest percentages of people who consider their health to be good are Lithuania, Latvia and Portugal, and those with high percentages are Ireland, Greece and Cyprus.
OB4 designates early leavers from education and training, and according to Figure 3b, the highest percentage of these people is in Romania, followed by Spain, Bulgaria and Italy, while the countries with a small percentage of the population for this indicator (under 4%) are Croatia, Greece, Ireland and Slovenia.As for Young people neither in employment nor in education and training, Italy, Romania and Bulgaria also have the highest percentages; for the first two countries, the values exceed 20%.The countries with low percentages of this indicator are the Netherlands, Sweden and Slovenia.Figure 3a describes the perception of the citizens of each European country regarding health (OB3) and corruption (OB16).Thus, according to the opinion of the population, the most corrupt countries are Bulgaria, Romania and Hungary, and the least corrupt are Finland, Denmark and Sweden.Regarding health perception, among the countries with the lowest percentages of people who consider their health to be good are Lithuania, Latvia and Portugal, and those with high percentages are Ireland, Greece and Cyprus.
OB4 designates early leavers from education and training, and according to Figure 3b, the highest percentage of these people is in Romania, followed by Spain, Bulgaria and Italy, while the countries with a small percentage of the population for this indicator (under 4%) are Croatia, Greece, Ireland and Slovenia.As for Young people neither in employment nor in education and training, Italy, Romania and Bulgaria also have the highest percentages; for the first two countries, the values exceed 20%.The countries with low percentages of this indicator are the Netherlands, Sweden and Slovenia.From Figure 4a, it can be observed that the country with the most positions held by women in senior management positions (board members) is France, where the percentage exceeds 45%, followed by Italy and the Nordic countries.The countries where this percentage does not exceed 10% are Cyprus, Hungary and Spain.For the income share of the bottom 40% of the population, it is observed that Italy and Romania have the highest From Figure 4a, it can be observed that the country with the most positions held by women in senior management positions (board members) is France, where the percentage exceeds 45%, followed by Italy and the Nordic countries.The countries where this percentage does not exceed 10% are Cyprus, Hungary and Spain.For the income share of the bottom 40% of the population, it is observed that Italy and Romania have the highest percentages for this indicator, exceeding 20%, and the countries with the lowest percentages are Sweden and The Netherlands.
Figure 4b shows that the Nordic countries (Finland, Sweden, Estonia) have the lowest exposure to PM2.5 air pollution, and the country with the highest exposure to polluted air is Croatia.For the population covered by the Covenant of Mayors for Climate & Energy signatories, Luxembourg has the lowest percentage of these people (less than 10%), and the highest percentage coverage is in Belgium, where the value exceeds 90%; most countries have values between 25% and 75%.
Sustainability 2024, 16, x FOR PEER REVIEW 8 of 20 percentages for this indicator, exceeding 20%, and the countries with the lowest percentages are Sweden and the Netherlands.Figure 4b shows that the Nordic countries (Finland, Sweden, Estonia) have the lowest exposure to PM2.5 air pollution, and the country with the highest exposure to polluted air is Croatia.For the population covered by the Covenant of Mayors for Climate & Energy signatories, Luxembourg has the lowest percentage of these people (less than 10%), and the highest percentage coverage is in Belgium, where the value exceeds 90%; most countries have values between 25% and 75%.According to Figure 5, Romania and Bulgaria are the only countries that have a percentage lower than 90% of people using at least basic sanitation services, most countries having values of these percentages that exceed 97.50%.Also, the country with the lowest percentage of forest area is Malta, and the countries with the largest forest area are the Nordic countries.Based on the data presented in Figure 6a, the Netherlands stands out as the country with the highest circular material use rate, surpassing 30%.In comparison, the majority of countries have values below 20% for this particular indicator.Interestingly, the Netherlands also boasts a higher percentage of high-speed internet coverage, exceeding 80%, which surpasses more than half of the countries in the European Union (EU).On the other According to Figure 5, Romania and Bulgaria are the only countries that have a percentage lower than 90% of people using at least basic sanitation services, most countries having values of these percentages that exceed 97.50%.Also, the country with the lowest percentage of forest area is Malta, and the countries with the largest forest area are the Nordic countries.
Sustainability 2024, 16, x FOR PEER REVIEW 8 of 20 percentages for this indicator, exceeding 20%, and the countries with the lowest percentages are Sweden and the Netherlands.Figure 4b shows that the Nordic countries (Finland, Sweden, Estonia) have the lowest exposure to PM2.5 air pollution, and the country with the highest exposure to polluted air is Croatia.For the population covered by the Covenant of Mayors for Climate & Energy signatories, Luxembourg has the lowest percentage of these people (less than 10%), and the highest percentage coverage is in Belgium, where the value exceeds 90%; most countries have values between 25% and 75%.According to Figure 5, Romania and Bulgaria are the only countries that have a percentage lower than 90% of people using at least basic sanitation services, most countries having values of these percentages that exceed 97.50%.Also, the country with the lowest percentage of forest area is Malta, and the countries with the largest forest area are the Nordic countries.Based on the data presented in Figure 6a, the Netherlands stands out as the country with the highest circular material use rate, surpassing 30%.In comparison, the majority of countries have values below 20% for this particular indicator.Interestingly, the Netherlands also boasts a higher percentage of high-speed internet coverage, exceeding 80%, which surpasses more than half of the countries in the European Union (EU).On the other Based on the data presented in Figure 6a, the Netherlands stands out as the country with the highest circular material use rate, surpassing 30%.In comparison, the majority of countries have values below 20% for this particular indicator.Interestingly, the Netherlands also boasts a higher percentage of high-speed internet coverage, exceeding 80%, which surpasses more than half of the countries in the European Union (EU).On the other hand, Greece exhibits the lowest percentage with approximately 20% high-speed internet coverage.In contrast, Malta has the highest value of 100% in this regard.
Figure 6b shows that most countries have a percentage of bathing sites with excellent water quality that exceeds 80%, Poland being the only country for which the percentage is below 50%.It should be noted that there are countries for which this indicator does not apply, such as Czechia, Luxembourg, Hungary, Austria and Slovakia.
Sustainability 2024, 16, x FOR PEER REVIEW 9 of 20 hand, Greece exhibits the lowest percentage with approximately 20% high-speed internet coverage.In contrast, Malta has the highest value of 100% in this regard.Figure 6b shows that most countries have a percentage of bathing sites with excellent water quality that exceeds 80%, Poland being the only country for which the percentage is below 50%.It should be noted that there are countries for which this indicator does not apply, such as Czechia, Luxembourg, Hungary, Austria and Slovakia.

Forecast
Since Romania's values presented in the European context were not among the most favorable, we wanted to identify the path this country is on in its journey towards achieving the SDGs.Therefore, based on available historical data, a forecast for the next three years was made.
Figure 7 presents the downward trend of the percentage of people at risk of poverty or social exclusion, forecasting, based on historical values, a continuous decrease in this indicator.Among the actions to be carried out within Romania's National Recovery and Resilience Plan for the European Pillar of Social Rights are reforms on the minimum income and voucher systems to encourage an increase in the number of women in the labor market [42].Thus, the number of people at risk of poverty or social exclusion should decrease, the hope being that this decrease will be a more significant one compared to the one achieved on the basis of historical data, therefore, of the policies that have been applied until now.

Forecast
Since Romania's values presented in the European context were not among the most favorable, we wanted to identify the path this country is on in its journey towards achieving the SDGs.Therefore, based on available historical data, a forecast for the next three years was made.
Figure 7 presents the downward trend of the percentage of people at risk of poverty or social exclusion, forecasting, based on historical values, a continuous decrease in this indicator.Among the actions to be carried out within Romania's National Recovery and Resilience Plan for the European Pillar of Social Rights are reforms on the minimum income and voucher systems to encourage an increase in the number of women in the labor market [42].Thus, the number of people at risk of poverty or social exclusion should decrease, the hope being that this decrease will be a more significant one compared to the one achieved on the basis of historical data, therefore, of the policies that have been applied until now.
According to Table 2, the model that had the lowest value for RMSE, i.e., ARIMA (0,1,1), has a Durbin-Watson test value less than 1, the errors being, in this case, positively autocorrelated, which can lead to an incorrect forecast.Thus, the method with the second lowest value of root mean square errors was chosen for which the value of the Durbin-Watson test is 1.78, close to 2. In this case, the errors are not autocorrelated, and the value of the Theil's U test is lower compared to 1; therefore, it can be stated that the forecasting technique is better than guessing.ing the SDGs.Therefore, based on available historical data, a foreca years was made.Figure 7 presents the downward trend of the percentage of peop or social exclusion, forecasting, based on historical values, a continu indicator.Among the actions to be carried out within Romania's Nat Resilience Plan for the European Pillar of Social Rights are reforms o come and voucher systems to encourage an increase in the number of market [42].Thus, the number of people at risk of poverty or social e crease, the hope being that this decrease will be a more significant o one achieved on the basis of historical data, therefore, of the policies plied until now.Figure 8 indicates a positive trend in the share of people perceiving their health as good or very good, and the forecast suggests that this trend will continue.Romania's National Recovery and Resilience Plan includes measures aimed at enhancing access to healthcare, particularly in rural areas, through the reform of the medical system [43].This increase in the percentage of the population perceiving their health positively could be attributed to multiple factors.Firstly, it may be a result of individuals feeling that the worst of the COVID-19 pandemic is behind them.Additionally, it could be influenced by the belief that the medical system will evolve to better serve their healthcare needs.
According to Table 2, the model that had the lowest value for RM (0,1,1), has a Durbin-Watson test value less than 1, the errors being, in th autocorrelated, which can lead to an incorrect forecast.Thus, the method lowest value of root mean square errors was chosen for which the valu Watson test is 1.78, close to 2. In this case, the errors are not autocorrelat of the Theil's U test is lower compared to 1; therefore, it can be stated th technique is better than guessing.Figure 8 indicates a positive trend in the share of people perceivin good or very good, and the forecast suggests that this trend will continu tional Recovery and Resilience Plan includes measures aimed at enh healthcare, particularly in rural areas, through the reform of the medical increase in the percentage of the population perceiving their health po attributed to multiple factors.Firstly, it may be a result of individual worst of the COVID-19 pandemic is behind them.Additionally, it could the belief that the medical system will evolve to better serve their health In Table 3, the annual increases can be observed, not exceeding one from one year to another.The model that best explains the historical d In Table 3, the annual increases can be observed, not exceeding one percentage point from one year to another.The model that best explains the historical data for making a forecast is double exponential smoothing for which the RMSE value is 0.73, the Durbin-Watson test is 1.82, very close to 2, and the Theil's U test has the lowest value, which is smaller compared to 1; it can be stated that the model is a better one than a simple guess.Regarding the fourth analyzed indicator, a noticeable downward trend is observed with the exception of the last year, which shows a slight increase-Figure 9.As desired, the trend for the forecast period is downward, in line with the first target of the fourth sustainable development goal entitled "Quality Education".
tainability 2024, 16, x FOR PEER REVIEW Regarding the fourth analyzed indicator, a noticeable downwar with the exception of the last year, which shows a slight increase-F the trend for the forecast period is downward, in line with the first sustainable development goal entitled "Quality Education".The predicted downward trend was identified using the non-seas method for which all the indicators described in Table 4 have valu correctness of the model.In this sense, the value of the Durbin-Wats which excludes the autocorrelation of lag 1 errors, and the value of lower compared to 1, indicating that the prediction made is better tha  The predicted downward trend was identified using the non-seasonal damped trend method for which all the indicators described in Table 4 have values that support the correctness of the model.In this sense, the value of the Durbin-Watson test is close to 2, which excludes the autocorrelation of lag 1 errors, and the value of the Theil's U test is lower compared to 1, indicating that the prediction made is better than guessing.
With a minor exception at the beginning of the analyzed period, where the trend initially showed a decline, the subsequent period demonstrates a consistent upward trend-Figure 10.The forecast obtained is optimistic, as it maintains a steady increase in values from one year to the next.According to the fifth target of the fifth objective of sustainable development, equal opportunities for obtaining management positions are desired [44] and, according to Romania's response to the European Commission regard-ing the National Recovery and Resilience Plan, an increase in the number of women in management positions at the level of state institutions [42].According to Table 5, the best model in terms of the value of root is ARIMA(2,1,1), but for this, the Durbin-Watson test value is lower makes the errors positively autocorrelated, a fact which could lead to Thus, the double exponential smoothing method was chosen for whic slightly higher compared to the previous case, but the value of the Du accepted; it can be stated that the errors are not significantly autocorr the forecast.According to Table 5, the best model in terms of the value of root mean square errors is ARIMA(2,1,1), but for this, the Durbin-Watson test value is lower compared to 1; this makes the errors positively autocorrelated, a fact which could lead to a wrong prediction.Thus, the double exponential smoothing method was chosen for which the RMSE value is slightly higher compared to the previous case, but the value of the Durbin-Watson test is accepted; it can be stated that the errors are not significantly autocorrelated to jeopardize the forecast.
For objective 12, the constant values from the beginning of the analyzed period are noted, followed by a decrease until 2019, a slight increase in 2020, followed by a further decrease in 2021-Figure 11.Based on the historical values and the consistent trend observed, the forecast aligns with expectations, as the data remain relatively stable and consistent.Romania's position among the countries with low values of this indicator highlights the need to enhance production efficiency and elevate the value of the indicator in order to facilitate the transition to a green economy.In this context, it can be noted that it is crucial to identify strategies and approaches that can optimize production processes and contribute to an overall improvement in this specific indicator.According to Table 6, the best method is double exponential sm value being 0.14 and the Durbin-Watson test having a value close to errors are not autocorrelated, while the Theil's U test value is less than  According to Table 6, the best method is double exponential smoothing, the RMSE value being 0.14 and the Durbin-Watson test having a value close to two, a sign that the errors are not autocorrelated, while the Theil's U test value is less than one.Target five of the 16th sustainable development goal is about reducing corruption; unfortunately, in people's perception, Romania is one of the most corrupt countries in the EU; therefore, it is very possible that, for this indicator, this country will not reach the desired target for 2030.At the end of the analyzed period, a slight increase in these values is observed, followed by a forecast in which the values increase moderately from one year to another; however, the increases should be higher, so the state should take measures to stop corruption-Figure 12.
Target five of the 16th sustainable development goal is about re unfortunately, in people's perception, Romania is one of the most corr EU; therefore, it is very possible that, for this indicator, this country desired target for 2030.At the end of the analyzed period, a slight incr is observed, followed by a forecast in which the values increase moder to another; however, the increases should be higher, so the state shou stop corruption-Figure 12.In the case of this indicator, the method for which the RMSE ha could not be chosen because the errors were negatively autocorrelated In the case of this indicator, the method for which the RMSE had a very low value could not be chosen because the errors were negatively autocorrelated; thus, the damped trend non-seasonal method was chosen for which the value of the Durbin-Watson test is much closer to 2 and the Theil's U test value is less compared to 1 (Table 7).To ensure success in achieving these goals, the connectivity and synergy between these goals must be taken into account.They should not be viewed individually but rather as a complex process by which the achievement of a goal leads to the achievement of progress in a related goal.As can be noted in the case of Table 8, where the correlations between the indicators are presented, the links between the analyzed targets are of medium to high intensity.Thus, the percentage of people at risk of poverty or social exclusion is positively correlated with the percentage of early leavers from education and training and negatively with the percentage of people with good or very good perceived health.

Discussion and Limitations
Forecasting the indicators of the Sustainable Development Goals is an element of interest among researchers because it is very possible that they will not be achieved, just like in the case of the Millennium Goals.The positive effect of the Millennium Goals on reducing poverty, diminishing gender equality, etc. should be noted, but part of this activity has been taken over by the SDGs [45,46].Although the lockdown caused by COVID-19 brought benefits to the environment [47,48], the pandemic period represented a turning point, being the first time a regression of these indicators was encountered [49,50].
Thus, the importance of identifying the balance between the economic, social and ecological domains is all the more obvious in the context of population growth and diminishing resources, researching energy consumption and renewable sources [51][52][53], pollution [54,55], poverty [56,57], education [58,59], etc.
Regarding the forecasted indicators, it can be observed that the situation would turn for the better; with all this, there are high risks that Romania will not reach the targets proposed by Agenda 2030.Thus, in terms of the percentage of people at risk of poverty or social exclusion, Romania has the highest value at the EU level, and the predicted decrease is, at most, moderate.Solutions to reduce this risk would be a reduction in taxes for people with low incomes; an increase in jobs, especially in rural areas; and the granting of social aid to people in need.The way in which Romanians perceive health seems to be inversely proportional to reality, the health system in this country being very poorly financed and lacking in staff [60].Although it seems that their perspective is in continuous improvement, it is desirable that this fact be supported by concrete actions, such as renovating hospitals, maybe even building new ones with European funds, and ensuring a sufficient number of medical personnel, especially in rural areas.And in terms of the school dropout rate, Romania is at the top of the European ranking, its predicted decrease being a very small one.As poverty and the need to work push parents not to send their children to school, the risk of dropping out could be significantly reduced if hot meals were provided at school and school materials were provided to children to facilitate this process of learning (it should be noted that there are already certain vouchers granted by the state in this regard).The percentage of women occupying senior management positions is also small, and the forecasted increase seems to be considerable; this fact can also be supported by Romania's National Recovery and Resilience Plan, which promotes an increase in this percentage in public institutions.Regarding the circular material use rate, Romania has the lowest values in the EU, and the forecast does not seem to be favorable.To mitigate this risk, it is recommended to deepen the legislative provisions and educate citizens and businesses in an ecological manner.And with regard to corruption, it can be observed that Romania is perceived as one of the most corrupt countries in the EU, and the forecast shows an improvement in this indicator, although the difference recorded is not very large compared to the last year of the analyzed period.In this case, the legislation must be improved to limit this phenomenon as much as possible.
The issue of early school dropout rates is a significant concern in Europe, particularly in Romania, where dropout rates are among the highest.According to Guerrero Puerta [61], educational policies aimed at promoting inclusion often inadvertently contribute to the exclusion of certain individuals.One proposed solution is to increase funding for support teachers or provide assistance to students with learning difficulties to help them catch up with their peers.Another challenge lies in the lack of school counselors who could provide support not only from an educational perspective but also emotionally.
The determinants of school dropout among young people from diverse backgrounds stem from various factors, including personal, socio-economic, educational and familial influences.These factors encompass a range of experiences in the social and school lives of young individuals, such as poor academic performance, strained relationships with teachers and personal or family issues.The long-term consequences of early school leaving extend to reduced incomes and increased unemployment rates among affected individuals, resulting in significant financial implications for both individuals and society as a whole.
High school dropout rates hinder economic growth, necessitate increased financial efforts to maintain employment levels and social cohesion and impede lifelong learning processes.Recommendations primarily focus on implementing specific policies and measures to support at-risk youth at both individual and group levels as well as within schools and areas recognized as high-risk for dropout (such as rural areas).
Additionally, it is crucial to prioritize improvements in the education system for the entire student population, as enhanced educational outcomes directly contribute to reducing school dropout rates.Furthermore, addressing these challenges requires a longterm perspective.Implementing a school-level time warning system facilitates the early identification of at-risk students, enabling the establishment of tailored monitoring and intervention plans to support their educational journey.
Another essential policy involves the reintegration of individuals who have previously dropped out of the education system but require qualifications to participate in the labor market effectively.By facilitating their return to education, we can mitigate the risk of poverty, a significant concern in Romania, where the country ranks poorly in EU indicators.
To further reduce this indicator, efforts should be made to combat discrimination against minorities and ensure the full integration of all individuals into the workforce.Specifically, initiatives aimed at empowering women, particularly those in rural areas, through training programs and employment opportunities, are essential.
Moreover, addressing health issues is paramount, as adopting a healthy lifestyle not only enhances citizens' productivity but also contributes to increased incomes.Romania, along with other EU member states, is committed to participating in the European Pillar of Social Rights Action Plan.In this regard, we recommend a comprehensive study of the 20 principles outlined in the pillar, focusing on identifying key principles and rights essential for creating fair and functional labor markets and welfare systems in 21st-century Europe.These principles should be adapted to the current conditions and state of affairs in Romania, with a primary goal of combating and reducing poverty and social exclusion.
Another critical indicator discussed in our current research, with significant implications for education and poverty, is corruption, an issue for which Romania ranks among the worst in the EU.While the eradication of corruption is an overarching goal, any progress made in this regard is invaluable.Therefore, it is essential to encourage individuals to openly discuss cases of corruption and report any instances they are aware of.
Furthermore, fostering collaboration among member states and ensuring the uniform adaptation of laws are vital recommendations.Adequate allocation of resources is imperative for the successful implementation of initiatives aimed at reducing corruption, in alignment with the National Anti-Corruption Strategy 2021-2025.However, implementing such initiatives faces challenges, including a lack of consistent political will and attitudes at the managerial level as well as legislative instability and insufficient support from Parliament.
In our assessment, these challenges stem from increasing political instability, particularly evident in the post-pandemic period, and electoral cycles.Newly elected government teams often introduce new legislative proposals, leading to delays in implementing programs and plans initiated by previous administrations.Therefore, we recommend several measures to address these issues:

•
Improving legislation to adapt to the challenges posed by the COVID-19 pandemic and the post-pandemic period; • Identifying and implementing best practices from previous iterations of the National Anti-Corruption Strategy;

•
Conducting impact studies on the implementation of new legislative acts aimed at addressing the business environment in our country.
Another significant indicator under discussion is the rate of circular material usage for which the prognosis appears discouraging, signaling the need to explore alternative solutions to enhance this rate.One contributing factor to the decreased use of materials compared to the past is their diminished quality and the widespread availability of new products.For instance, whenever a new phone model is released, there is often a desire among consumers to acquire it, leading to a surplus of unused devices at home.While some companies specialize in refurbishing phones, the demand for very old models may be limited.However, the general population is often unaware that valuable raw materials can be extracted from older phones for use in new products.
Therefore, current recommendations focus on developing sustainable materials that are reusable or repairable.Additionally, efforts should be made to educate the public about recycling methods, especially for products that cannot be easily repurposed.Despite the implementation of economic (such as specific taxes), coercive (legislative) and administrative measures aimed at increasing recycling rates, significant improvements have not yet been realized.
To mitigate the risk of failing to meet the indicator by 2030, several actions are recommended:

•
Conducting comprehensive documentary research to analyze the current situation in Romania;

•
Identifying best practices within the EU regarding strategies and governance in the circular economy field; • Updating and improving legislation, particularly laws specific to the circular economy; • Educating citizens and businesses about ecological practices, given the relatively low level of involvement of Romanian citizens in circular economy activities.
While some authors have sought to assess the achievement of Sustainable Development Goals (SDGs) on a global or continental scale [15,16], our study aimed to examine this within a narrower context.Romania was selected due to its portrayal of some of the poorest results within the EU.Additionally, given the complexity of the topic and the extensive array of indicators involved, the authors opted to focus solely on specific goals rather than discussing all of them collectively within a single article [30,31,62].
It is important to acknowledge that the limited time interval poses a challenge in developing models that provide a more accurate representation of reality.However, considering that the 2030 Agenda was signed in 2015, the available years for analysis correspond to the period after its implementation.For most indicators, the data available extend until 2021 with only two exceptions where the data for 2020 were used to describe the temporal evolution without the ability to generate forecasts due to the limited number of observed years.
Also, due to the small number of years, superior data forecasting methods could not be used because the number of years required by them is higher than that provided by Eurostat.However, it can be noted, based on Tables 2-7, that those models were chosen for which the RMSE has the minimum value to ensure the smallest possible differences between the predicted and actual values.Also, for all the estimated models, it was found that the values of the other tests, namely Durbin-Watson and Theil's U, always respected the methodological indications to consider the provided models satisfactory.Thus, in the case of the forecasts made, the value of the Durbin-Watson test was always close to 2, which shows an independence of the data from the past errors, while for Theil's U test, the values obtained were below the threshold of 1, thus suggesting that the prediction provided is better than guessing.
The forecasting for the other indicators was conducted using the autoregressive integrated moving average (ARIMA) method, which can provide more insight on the multiple facets of the sustainable development in the EU context [63].

Conclusions
The dedicated section on the European context highlights Romania's concerning standings across various indicators.Notably, Romania ranks highest in the percentage of people at risk of poverty or social exclusion.Moreover, it ranks second lowest in government investment in agricultural research and development and displays a meager number of patent applications to the European Patent Office.Corruption perception places Romania among the top three most corrupt EU countries, while the perceived health of its citizens contrasts with the performance of the Romanian medical system.
Education outcomes in Romania are also lackluster with the highest percentage of early school leavers and the second highest percentage of young individuals not engaged in employment or education and training.Furthermore, Romania struggles with gender representation in senior management positions but exhibits a relatively higher income share among the bottom 40% of its population.
Basic sanitation service utilization in Romania remains notably low, and the country exhibits the lowest circular material use in the EU.Given these indicators' depiction of Romania's challenges, this study aims to forecast potential pathways forward based on historical data provided by Eurostat.Based on the forecasts derived from historical values, the number of individuals at risk of poverty or social exclusion is expected to decrease.Additionally, there should be an increase in the percentage of people who perceive their health as good.Simultaneously, there may be a slight decline in the percentage of early leavers from education and training over the projected three-year period.An increase in the number of positions held by women in senior management positions (board members) is also expected, while the indicator for corruption also tends to increase, a higher value being attributed to a lower degree of corruption.

Figure 1 .Figure 1 .
Figure 1.Visualization of countries from the perspective of OB1 and OB7.According to Figure 2, Denmark is the country that stands out positively for both analyzed objectives, having a value of over 25 euros per inhabitant provided by the government to support agricultural research and development.At the other pole are Poland and Romania with the lowest values for both indicators (0.10 euro per inhabitant and 16.72 per million inhabitants for Poland, respectively 1.10 euro and 2.40 patents per million inhabitants for Romania).

Figure 2 .
Figure 2. Visualization of countries from the perspective of OB2 and OB9.

Figure 2 .
Figure 2. Visualization of countries from the perspective of OB2 and OB9.

Sustainability 2024 , 20 Figure 2 .
Figure 2. Visualization of countries from the perspective of OB2 and OB9.

Figure 5 .
Figure 5. Visualization of countries from the perspective of OB6 and OB15.

Figure 5 .
Figure 5. Visualization of countries from the perspective of OB6 and OB15.

Figure 5 .
Figure 5. Visualization of countries from the perspective of OB6 and OB15.

Table 4 .
[42]forecast results., equal opportunities for obtaining management positio and, according to Romania's response to the European Commission tional Recovery and Resilience Plan, an increase in the number of wom positions at the level of state institutions[42]. development
Target five of the 16th sustainable development goal is about re