arXiv subjects

Stefan Thurner

Publications and source records attributed to Stefan Thurner.

At least 37 records · Page 2Linked to original sources

arXiv2023

Forensic analysis of the Turkey 2023 presidential election reveals extreme vote swings in remote areas

Concerns about the integrity of Turkey's elections have increased with the recent transition from a parliamentary democracy to an executive presidency under Recep Tayyip Erdogan. Election forensics tools are used to identify statistical traces of certain types of electoral fraud, providing important information about the integrity and validity of democratic elections. Such analyses of the 2017 and 2018 Turkish elections revealed that malpractices such as ballot stuffing or voter manipulation may indeed have played a significant role in determining the election results. Here, we apply election forensic statistical tests for ballot stuffing and voter manipulation to the results of the 2023 presidential election in Turkey. We find that both rounds of the 2023 presidential election exhibit similar statistical irregularities to those observed in the 2018 presidential election, however the magnitude of these distortions has decreased. We estimate that 2.4% (SD 1.9%) and 1.9% (SD 1.7%) of electoral units may have been affected by ballot stuffing practices in favour of Erdogan in the first and second rounds, respectively, compared to 8.5% (SD 3.9%) in 2018. Areas with smaller polling stations and fewer ballot boxes had significantly inflated votes and turnout, again, in favor of Erdogan. Furthermore, electoral districts with two or fewer ballot boxes were more likely to show large swings in vote shares in favour of Erdogan from the first to the second round. Based on a statistical model, we estimate that these swings translate into 342,000 excess votes (SD 4,900) or 0.64% for Erdogan. Our results suggest that Turkish elections continue to be riddled with statistical irregularities, that may be indicative of electoral fraud.

stat.AP↗

arXiv2023

Unraveling cradle-to-grave disease trajectories from multilayer comorbidity networks

We aim to comprehensively identify typical life-spanning trajectories and critical events that impact patients' hospital utilization and mortality. We use a unique dataset containing 44 million records of almost all inpatient stays from 2003 to 2014 in Austria to investigate disease trajectories. We develop a new, multilayer disease network approach to quantitatively analyse how cooccurrences of two or more diagnoses form and evolve over the life course of patients. Nodes represent diagnoses in age groups of ten years; each age group makes up a layer of the comorbidity multilayer network. Inter-layer links encode a significant correlation between diagnoses (p $<$ 0.001, relative risk $>$ 1.5), while intra-layers links encode correlations between diagnoses across different age groups. We use an unsupervised clustering algorithm for detecting typical disease trajectories as overlapping clusters in the multilayer comorbidity network. We identify critical events in a patient's career as points where initially overlapping trajectories start to diverge towards different states. We identified 1,260 distinct disease trajectories (618 for females, 642 for males) that on average contain 9 (IQR 2-6) different diagnoses that cover over up to 70 years (mean 23 years). We found 70 pairs of diverging trajectories that share some diagnoses at younger ages but develop into markedly different groups of diagnoses at older ages. The disease trajectory framework can help us to identify critical events as specific combinations of risk factors that put patients at high risk for different diagnoses decades later. Our findings enable a data-driven integration of personalized life-course perspectives into clinical decision-making.

physics.med-ph↗

arXiv2023

Shock propagation from the Russia-Ukraine conflict on international multilayer food production network determines global food availability

Dependencies in the global food production network can lead to shortages in numerous regions, as demonstrated by the impacts of the Russia-Ukraine conflict on global food supplies. Here, we reveal the losses of $125$ food products after a localized shock to agricultural production in $192$ countries and territories using a multilayer network model of trade (direct) and conversion of food products (indirect), thereby quantifying $10^8$ shock transmissions. We find that a complete agricultural production loss in Ukraine has heterogeneous impacts on other countries, causing relative losses of up to $89\%$ in sunflower oil and $85\%$ in maize via direct effects, and up to $25\%$ in poultry meat via indirect impacts. Whilst previous studies often treated products in isolation and did not account for product conversion during production, our model studies the global propagation of local supply shocks along both production and trade relations, allowing comparison of different response strategies.

econ.GN↗

arXiv2023

Estimating the impact of supply chain network contagion on financial stability

Realistic credit risk assessment, the estimation of losses from counterparty's failure, is central for the financial stability. Credit risk models focus on the financial conditions of borrowers and only marginally consider other risks from the real economy, supply chains in particular. Recent pandemics, geopolitical instabilities, and natural disasters demonstrated that supply chain shocks do contribute to large financial losses. Based on a unique nation-wide micro-dataset, containing practically all supply chain relations of all Hungarian firms, together with their bank loans, we estimate how firm-failures affect the supply chain network, leading to potentially additional firm defaults and additional financial losses. Within a multi-layer network framework we define a financial systemic risk index (FSRI) for every firm, quantifying these expected financial losses caused by its own- and all the secondary defaulting loans caused by supply chain network (SCN) shock propagation. We find a small fraction of firms carrying substantial financial systemic risk, affecting up to 16% of the banking system's overall equity. These losses are predominantly caused by SCN contagion. For every bank we calculate the expected loss (EL), value at risk (VaR) and expected shortfall (ES), with and without accounting for SCN contagion. We find that SCN contagion amplifies the EL, VaR, and ES by a factor of 4.3, 4.5, and 3.2, respectively. These findings indicate that for a more complete picture of financial stability and realistic credit risk assessment, SCN contagion needs to be considered. This newly quantified contagion channel is of potential relevance for regulators' future systemic risk assessments.

q-fin.ST↗

arXiv2023

Estimating the loss of economic predictability from aggregating firm-level production networks

To estimate the reaction of economies to political interventions or external disturbances, input-output (IO) tables -- constructed by aggregating data into industrial sectors -- are extensively used. However, economic growth, robustness, and resilience crucially depend on the detailed structure of non-aggregated firm-level production networks (FPNs). Due to non-availability of data little is known about how much aggregated sector-based and detailed firm-level-based model-predictions differ. Using a nearly complete nationwide FPN, containing 243,399 Hungarian firms with 1,104,141 supplier-buyer-relations we self-consistently compare production losses on the aggregated industry-level production network (IPN) and the granular FPN. For this we model the propagation of shocks of the same size on both, the IPN and FPN, where the latter captures relevant heterogeneities within industries. In a COVID-19 inspired scenario we model the shock based on detailed firm-level data during the early pandemic. We find that using IPNs instead of FPNs leads to errors up to 37% in the estimation of economic losses, demonstrating a natural limitation of industry-level IO-models in predicting economic outcomes. We ascribe the large discrepancy to the significant heterogeneity of firms within industries: we find that firms within one sector only sell 23.5% to and buy 19.3% from the same industries on average, emphasizing the strong limitations of industrial sectors for representing the firms they include. Similar error-levels are expected when estimating economic growth, CO2 emissions, and the impact of policy interventions with industry-level IO models. Granular data is key for reasonable predictions of dynamical economic systems.

econ.GN↗

arXiv2023

A parameter-free population-dynamical approach to health workforce supply forecasting of EU countries

Many countries faced challenges in their health workforce supply like impending retirement waves, negative population growth, or a suboptimal distribution of resources across medical sectors even before the pandemic struck. Current quantitative models are often of limited usability as they either require extensive individual-level data to be properly calibrated or (in the absence of such data) become too simplistic to capture key demographic changes or disruptive epidemiological shocks like the SARS-CoV-2 pandemic. We propose a novel population-dynamical and stock-flow-consistent approach to health workforce supply forecasting that is complex enough to address dynamically changing behaviors while requiring only publicly available timeseries data for complete calibration. We demonstrate the usefulness of this model by applying it to 21 European countries to forecast the supply of generalist and specialist physicians until 2040, as well as how Covid-related mortality and increased healthcare utilization might impact this supply. Compared to staffing levels required to keep the physician density constant at 2019 levels, we find that in many countries there is indeed a significant trend toward decreasing density for generalist physicians at the expense of increasing densities for specialists. The trends for specialists are exacerbated in many Southern and Eastern European countries by expectations of negative population growth. Compared to the expected demographic changes in the population and the health workforce, we expect a limited impact of Covid on these trends even under conservative modelling assumptions. It is of the utmost importance to devise tools for decision makers to influence the allocation and supply of physicians across fields and sectors to combat these imbalances.

stat.AP↗

arXiv2022

Evidence-based policy-making in sports funding using a data-driven optimization approach

Many European countries face rising obesity rates among children, compounded by decreased opportunities for sports activities during the COVID-19 pandemic. Access to sports facilities depends on multiple factors, such as geographic location, proximity to population centers, budgetary constraints, and other socio-economic covariates. Here we show how an optimal allocation of government funds towards sports facilitators (e.g. sports clubs) can be achieved in a data-driven simulation model that maximizes children's access to sports facilities. We compile a dataset for all 1,854 football clubs in Austria, including estimates for their budget, geolocation, tally, and the age profile of their members. We find a characteristic sub-linear relationship between the number of active club members and the budget, which depends on the socio-economic conditions of the clubs' municipality. In the model, where we assume this relationship to be causal, we evaluate different funding strategies. We show that an optimization strategy where funds are distributed based on regional socio-economic characteristics and club budgets outperforms a naive approach by up to 117\% in attracting children to sports clubs for 5 million Euros of additional funding. Our results suggest that the impact of public funding strategies can be substantially increased by tailoring them to regional socio-economic characteristics in an evidence-based and individualized way.

physics.soc-ph↗

arXiv2022

To what extent homophily and influencer networks explain song popularity

Forecasting the popularity of new songs has become a standard practice in the music industry and provides a comparative advantage for those that do it well. Considerable efforts were put into machine learning prediction models for that purpose. It is known that in these models, relevant predictive parameters include intrinsic lyrical and acoustic characteristics, extrinsic factors (e.g., publisher influence and support), and the previous popularity of the artists. Much less attention was given to the social components of the spreading of song popularity. Recently, evidence for musical homophily - the tendency that people who are socially linked also share musical tastes - was reported. Here we determine how musical homophily can be used to predict song popularity. The study is based on an extensive dataset from the last.fm online music platform from which we can extract social links between listeners and their listening patterns. To quantify the importance of networks in the spreading of songs that eventually determines their popularity, we use musical homophily to design a predictive influence parameter and show that its inclusion in state-of-the-art machine learning models enhances predictions of song popularity. The influence parameter improves the prediction precision (TP/(TP+FN)) by about 50% from 0.14 to 0.21, indicating that the social component in the spreading of music plays at least as significant a role as the artist's popularity or the impact of the genre.

physics.soc-ph↗

arXiv2022

Evolving Collaboration, Dependencies, and Use in the Rust Open Source Software Ecosystem

Open-source software (OSS) is widely spread in industry, research, and government. OSS represents an effective development model because it harnesses the decentralized efforts of many developers in a way that scales. As OSS developers work independently on interdependent modules, they create a larger cohesive whole in the form of an ecosystem, leaving traces of their contributions and collaborations. Data harvested from these traces enable the study of large-scale decentralized collaborative work. We present curated data on the activity of tens of thousands of developers in the Rust ecosystem and the evolving dependencies between their libraries. The data covers seven years of developer contributions to Rust libraries and can be used to reconstruct the ecosystem's development history, such as growing developer collaboration networks or dependency networks. These are complemented by statistics on downloads and popularity, tracking the dynamics of use and success over time. Altogether the data give a comprehensive view of several dimensions of the ecosystem.

cs.SE↗

arXiv2022

Homophily-based social group formation in a spin-glass self-assembly framework

Homophily, the tendency of humans to attract each other when sharing similar features, traits, or opinions has been identified as one of the main driving forces behind the formation of structured societies. Here we ask to what extent homophily can explain the formation of social groups, particularly their size distribution. We propose a spin-glass-inspired framework of self-assembly, where opinions are represented as multidimensional spins that dynamically self-assemble into groups; individuals within a group tend to share similar opinions (intra-group homophily), and opinions between individuals belonging to different groups tend to be different (inter-group heterophily). We compute the associated non-trivial phase diagram by solving a self-consistency equation for 'magnetization' (combined average opinion). Below a critical temperature, there exist two stable phases: one ordered with non-zero magnetization and large clusters, the other disordered with zero magnetization and no clusters. The system exhibits a first-order transition to the disordered phase. We analytically derive the group-size distribution that successfully matches empirical group-size distributions from online communities.

cond-mat.stat-mech↗

arXiv2022

Small coverage effect in epidemic network models shows that masks can become more effective with less people wearing them

The effectiveness of non-pharmaceutical interventions to curb the spread of SARS-CoV-2 is determined by numerous contextual factors, including adherence. Conventional wisdom holds that the effectiveness of protective behaviour such as wearing masks always increases with the number of people adopting it. Here we show in a simulation study that this is not true in general. We employ a parsimonious network model based on the well-established empirical facts that (i) adherence to such interventions wanes over time and (ii) individuals tend to align their adoption strategies with their close social ties (homophily). When combining these assumptions, a broad dynamical regime emerges where the individual-level infection risk reduction for those adopting protective behaviour increases as the adherence to protective behavior decreases. For instance, for a protective coverage of 10% we find the infection risk for adopting individuals can be reduced by close to 30% compared to situations where the coverage is 60%. Using estimates for the effectiveness of surgical masks, we find that reductions in relative risk of masking versus non-masking individuals range between 5% and 15%, i.e., vary by a factor of three. This small coverage effect originates from system-dynamical network properties that conspire to increase the chance that an outbreak will be over before the pathogen is able to invade small but tightly connected groups of individuals that protect themselves. Our results contradict the popular belief that masking becomes ineffectual as more people drop their masks and might have far-reaching implications for the protection of vulnerable population groups under resurgent infection waves.

physics.soc-ph↗

arXiv2022

Equivalence of information production and generalized entropies in complex processes

Complex systems that are characterized by strong correlations and fat-tailed distribution functions have been argued to be incompatible within the framework of Boltzmann-Gibbs entropy. As an alternative, so-called generalized entropies were proposed and intensively studied. Here we show that this incompatibility is a misconception. For a broad class of processes, Boltzmann entropy the log multiplicity remains the valid entropy concept, however, for non-i.i.d., non-multinomial, and non-ergodic processes, Boltzmann entropy is not of Shannon form. The correct form of Boltzmann entropy can be shown to be identical with generalized entropies. We derive this result for all processes that can be mapped reversibly to adjoint representations where processes are i.i.d.. In these representations the information production is given by the Shannon entropy. We proof that over the original sampling space this yields functionals that are identical to generalized entropies. The problem of constructing adequate context-sensitive entropy functionals therefore can be translated into the much simpler problem of finding adjoint representations. The method provides a comprehensive framework for a statistical physics of strongly correlated systems and complex processes.

cond-mat.stat-mech↗

arXiv2022

Specialization in Criminal Careers

We use a comprehensive longitudinal dataset on criminal acts over five years in a European country to study specialization in criminal careers. We cluster crime categories by their relative co-occurrence within criminal careers, deriving a natural, data-based taxonomy of criminal specialization. Defining specialists as active criminals who stay within one category of offending behavior, we study their socio-demographic attributes, geographic range, and positions in their collaboration networks, relative to their generalist counterparts. In comparison to generalists, specialists tend to be older, more likely to be female, operate within a smaller geographic range, and collaborate in smaller, more tightly-knit local networks. We observe that specialists are more intensely embedded in criminal networks and find evidence that specialization indeed reflects division of labor and organization.

physics.soc-ph↗

arXiv2022

Stress-testing the Resilience of the Austrian Healthcare System Using Agent-Based Simulation

Patients do not access physicians at random but rather via naturally emerging networks of patient flows between them. As retirements, mass quarantines and absence due to sickness during pandemics, or other shocks thin out these networks, the system might be pushed closer to a tipping point where it loses its ability to deliver care to the population. Here we propose a data-driven framework to quantify the regional resilience to such shocks of primary and secondary care in Austria via an agent-based model. For each region and medical specialty we construct detailed patient-sharing networks from administrative data and stress-test these networks by removing increasing numbers of physicians from the system. This allows us to measure regional resilience indicators describing how many physicians can be removed from a certain area before individual patients won't be treated anymore. We find that such tipping points do indeed exist and that regions and medical specialties differ substantially in their resilience. These systemic differences can be related to indicators for individual physicians by quantifying how much their hypothetical removal would stress the system (risk score) or how much of the stress from the removal of other physicians they would be able to absorb (benefit score). Our stress-testing framework could enable health authorities to rapidly identify bottlenecks in access to care as well as to inspect these naturally emerging physician networks and how potential absences would impact them.

physics.soc-ph↗

arXiv2022

Loss of sustainability in scientific work

For decades the number of scientific publications has been rapidly increasing, effectively out-dating knowledge at a tremendous rate. Only few scientific milestones remain relevant and continuously attract citations. Here we quantify how long scientific work remains being utilized, how long it takes before today's work is forgotten, and how milestone papers differ from those forgotten. To answer these questions, we study the complete temporal citation network of all American Physical Society journals. We quantify the probability of attracting citations for individual publications based on age and the number of citations they have received in the past. We capture both aspects, the forgetting and the tendency to cite already popular works, in a microscopic generative model for the dynamics of scientific citation networks. We find that the probability of citing a specific paper declines with age as a power law with an exponent of $α\sim -1.4$. Whenever a paper in its early years can be characterized by a scaling exponent above a critical value, $α_c$, the paper is likely to become "ever-lasting". We validate the model with out-of-sample predictions, with an accuracy of up to 90% (AUC $\sim 0.9$). The model also allows us to estimate an expected citation landscape of the future, predicting that 95% of papers cited in 2050 have yet to be published. The exponential growth of articles, combined with a power-law type of forgetting and papers receiving fewer and fewer citations on average, suggests a worrying tendency toward information overload and raises concerns about scientific publishing's long-term sustainability.

physics.soc-ph↗

arXiv2022

Propagation of disruptions in supply networks of essential goods: A population-centered perspective of systemic risk

The Covid-19 pandemic drastically emphasized the fragility of national and international supply networks (SNs),leading to significant supply shortages of essential goods for people, such as food and medical equipment. Severe disruptions that propagate along complex SNs can expose the population of entire regions or even countries to these risks. A lack of both, data and quantitative methodology, has hitherto hindered us to empirically quantify the vulnerability of the population to disruptions. Here we develop a data-driven simulation methodology to locally quantify actual supply losses for the population that result from the cascading of supply disruptions. We demonstrate the method on a large food SN of a European country including 22,938 business premises, 44,355 supply links and 116 local administrative districts. We rank the business premises with respect to their criticality for the districts' population with the proposed systemic risk index, SRIcrit, to identify around 30 premises that -- in case of their failure -- are expected to cause critical supply shortages in sizable fractions of the population. The new methodology is immediately policy relevant as a fact-driven and generalizable crisis management tool. This work represents a starting point for quantitatively studying SN disruptions focused on the well-being of the population.

econ.GN↗

arXiv2021

Inequality in economic shock exposures across the global firm-level supply network

For centuries, national economies created wealth by engaging in international trade and production. The resulting international supply networks not only increase wealth for countries, but also create systemic risk: economic shocks, triggered by company failures in one country, may propagate to other countries. Using global supply network data on the firm-level, we present a method to estimate a country's exposure to direct and indirect economic losses caused by the failure of a company in another country. We show the network of systemic risk-flows across the world. We find that rich countries expose poor countries much more to systemic risk than the other way round. We demonstrate that higher systemic risk levels are not compensated with a risk premium in GDP, nor do they correlate with economic growth. Systemic risk around the globe appears to be distributed more unequally than wealth. These findings put the often praised benefits for developing countries from globalized production in a new light, since they relate them to the involved risks in the production processes. Exposure risks present a new dimension of global inequality, that most affects the poor in supply shock crises. It becomes fully quantifiable with the proposed method.

econ.GN↗

arXiv2021

Bali's ancient rice terraces: A Hamiltonian approach

We propose a Hamiltonian approach to reproduce the relevant elements of the centuries-old Subak irrigation system in Bali, showing a cluster-size power-law distribution. The resulting system presents two equilibria. The first balance is between energy and entropy contributions. The second lies in the energy contribution: a local Potts-type interaction and a long-range anti-ferromagnetic one without attenuation. Finite-size scaling analysis shows that as a result of the second balance, the transition balancing energy and entropy contributions for low values of the antiferromagnetic contribution is absorbed by the transition for high values of that contribution as the system size increases. The system thus presents only a weak-first order phase transition at the thermodynamic limit. This extends the Hamiltonian framework to a new domain of coupled human-environmental interactions.

cond-mat.stat-mech↗