The global financial crisis hit the Serbian economy severely in late 2008. The subsequent decade has been characterized by negative or very modest economic growth and Serbia is now just slightly above the development level of ten years ago. This paper analyses the most important economic milestones during this decade and investigates why only modest progress has been made, despite various measures implemented by the Serbian government. It examines the background to Serbia?s delayed transition and analyses the effects of the global economic crisis on the Serbian economy. It outlines the policy responses and their results, focusing on public finance, foreign trade, reindustrialisation, FDI, the labour market, and sources of growth. The paper sets out the key challenges to accelerating Serbia?s economic growth and identifies the main elements of a new long-term development strategy.

Informalities in the Serbian economy seem to be relatively persistent and widespread, as suggested by some empirical evidence and numerous anecdotal evidence provided by citizens and companies. Nevertheless, the shadow—or informal—economy is a rather complicated phenomenon with a complex set of causes and consequences, implying that there are no universally accepted definitions nor quantification methods for determining the relative size of its impact. This paper takes a pioneering approach using the “monetary method” or the “currency demand approach” for the Serbian economy, based on econometric estimates of the demand for money. In this context, we rely on the well-known currency demand approach, expanding and applying it to the case of Serbia. Against the backdrop of the key research question, we found that the shadow economy in Serbia has decreased to roughly 20% of GDP in late 2021 from roughly 30% in 2005.

This work explores the relationship between water quality and public trust in their water supply, in an arsenic-affected region of Serbia. The results from an online public survey are presented and subjected to Pearson’s correlation, cluster analysis, and principal component analysis. In general, survey respondents in settlements with known arsenic issues had a poor opinion on the quality of their tap water. This poor opinion was underlined by their consumption of bottled water, with more than 43% of responders purchasing at least 5 L of bottled water a week. In addition to the great economic cost, the relatively low plastic recycling rate in Serbia means that this also has a very negative effect on the environment, as most of the plastic bottles are sent to landfill, to degrade slowly into microplastics, whilst leaching a variety of chemical contaminants into the surroundings. In the area studied, the poor public opinion of the water quality is a realistic reflection of water at the tap. Although significant, the additional environmental pollution from bottled water consumption should nonetheless be of secondary consideration in comparison to the health risks associated with chemical contaminants in the study area, with local waterworks requiring significant financial assistance if they are to meaningfully improve tap water quality.

Big data methods are becoming an important tool for tax fraud detection around the world. Unsupervised learning approach is the dominant framework due to the lack of label and ground truth in corresponding data sets although these methods suffer from low interpretability. HUNOD, a novel hybrid unsupervised outlier detection method for tax evasion risk management, is presented in this paper. In contrast to previous methods proposed in the literature, the HUNOD method combines two outlier detection approaches based on two different machine learning designs (i.e, clustering and representational learning) to detect and internally validate outliers in a given tax dataset. The HUNOD method allows its users to incorporate relevant domain knowledge into both constituent outlier detection approaches in order to detect outliers relevant for a given economic context. The interpretability of obtained outliers is achieved by training explainable-by-design surrogate models over results of unsupervised outlier detection methods. The experimental evaluation of the HUNOD method is conducted on two datasets derived from the database on individual personal income tax declarations collected by the Tax Administration of Serbia. The obtained results show that the method indicates between 90% and 98% internally validated outliers depending on the clustering configuration and employed regularization mechanisms for representational learning.