Solid background in probability theory and stochastic processes (knowledge about stochastic differential equations is a plus); good knowledge of statistics or data science (Bayesian statistics is a plus); programming skills, preferably in R or Python.
The project develops and implements statistical methods for diffusion processes embedded in state-space formulations, specifically (i) diffusion processes that are subject to unobserved switches between discrete states (i.e. regime-switching diffusion processes), and (ii) diffusion processes that are only indirectly observed via a measurement process, such that the diffusion process is the state process within a continuous-time state-space model.
(i) development of simulation-based Bayesian methods as well as approximate likelihood-based frequentist methods for inference in state-space diffusion processes, (ii) demonstration of the feasibility of the new methods in case studies using real data from (innovation) economics, ecology, the environment and medicine, (iii) development of user-friendly software for the dissemination of the methods to practitioners.
Strong background in statistics and data science. Solid background in multivariate stochastic processes/time series is a plus. Good programming skills, preferably in R or Python.
Complex data sets needed to understand climate evolution and economic policies consist of many time-series of different nature and spatio-temporal resolutions. The usual approach is therefore to analyse each time-series separately, but this is an inefficient use of the data, and since some (or all) of these time-series are strongly correlated, this might lead to wrong inferences. Furthermore, it is not clear how results from single analyses should be combined to an overall answer. The aim of this project is the development of models with multivariate responses that allow for complex dependence structures and latent variables, to fully exploit the richness of the data and make meaningful inference. A natural model class is dynamic structural equation models with latent variables, which can be used to study the evolution of observed and latent variables as well as the structural equation models over time.
(i) Development of a model framework for multi-dimensional dynamic data, including covariate effects, dependence structures and latent variables; (ii) Development of methods for detection of changes in the distribution of the data; (iii) Development of methods to detect the time order in which changes happen for different variables.
Desired Profile of Qualification: Strong background in statistics and data science. Solid background in multivariate stochastic processes/time series is a plus. Good programming skills, preferably in R or Python.
The project develops statistical method for data consisting of time-series of functional data, e.g., daily observations of curves and surfaces. Methods from functional data analysis will be combined with methods from classical time- series analysis in order to model variation within functional observations as well as dependence over time, and carry out valid statistical inference.
(i) Development of model framework for one-dimensional time-series of functional data, including covariate effects and dependence over time, and methods for hypothesis tests; (ii) Development of methods for change point detection, i.e. detection of changes in the distribution of the functional data; (iii) Extension of the framework for multi-dimensional data.
Strong background in dynamic economic modelling (ideally also in the area of agent-based simulations). Solid background in Statistics or Data Science is a plus.
The objective of this project is to develop computationally feasible and at the same time rigorous approaches for estimating parameters in mid-sized or even large agent-based models in economics. The existing literature so far has applied methods, like Bayesian estimation, mainly in the context of simple agent-based models. Increasing the size and the complexity of the considered model induces an increase of the dimensionality of the relevant parameter space as well as potential issues of non-ergodicity and extensive run times of the model. In this project expertise in efficient Bayesian estimation of models without (closed) likelihood functions will be combined with experience in designing and estimating industry- and macro-level agent-based models.
(i) Development of computationally feasible approaches for Bayesian estimation of mid-sized and large agent-based models; (ii) Examination of theoretical properties of the developed estimation scheme; (iii) proof of concept of the applicability of the developed approach using established agent-based models like the Eurace@Unibi or the CATS model.
The candidate should have a solid quantitative knowledge in statistics or data science and experience with dynamic economic modeling.
The aim of this project is to investigate how the introduction of machine learning techniques (e.g. Bayesian learning, random forests, k-Nearest-Neighbours), in expectation formation affects the stability of macro-financial systems. Studies of agent-based models with heterogeneous expectations show that the interaction between costly sophisticated (rational) rules and freely available rules of thumb may lead to complex dynamics. This project is, to the best of our knowledge, the first to systematically study the role of machine learning in heterogeneous expectations frameworks.
The project is expected to provide fundamental new insights into whether machine learning can stabilise complex dynamics and/or whether information gathering costs may be destabilising when machine learning tools are available to the agents. The effects of machine learning on the dynamics will be assessed in a range of frameworks, such as cobweb, asset pricing and macro-economic models.
Strong quantitative background in Economics or Management Sciences. Solid knowledge in Computational Social Sciences is a bonus.
When people lack information, they build mental models reflecting interdependencies among variables that are not always correctly specified or shared, see Martignoni et al. (2016). The ESR will simulate models (i.e., convenient cognitive simplification of complex conditions) in the presence of under- or over-specification and assess the practical and managerial implications of the “errors” in the model. The investigation is then extended to allow for multiple agents who routinely interact, exchanging information and beliefs, but these exchanges seldom lead to complete agreement.
The ESR will (i) build insights on the effect of (diverse) models’ misspecification, using analytical or computational models; (ii) analyse models of "disruption and repair", describing situations in which some initial compatibility is broken, diverse world views emerge and new ways to "cooperate" (or jointly interpret the world) must be bargained.
Strong analytical skills, strong interest in economic modelling, ability to work on both theoretical and empirical problems. Interest in climate policy issues.
The aim of the project is to provide micro-economic foundations for the assignment of liabilities for carbon emissions in supply chains and correlatively to identify actors that are most exposed to climate transition risks. In this perspective, we will develop axiomatic models of the allocation of externalities in production networks.
The project will put forward transparent principles for the allocation of carbon emissions in global supply chains. Such principles could be of particular interest for regulators in the field of climate policy. They will in particular allow to overcome current issues on "double-counting" of emissions and, more broadly, help to determine the liability of economics actors for anthropogenic climate change. From a more quantitative perspective, the project will allow to develop metrics to quantify the exposure of economic and financial actors to climate "transition" risk. Such metrics could be in particular of interest for the financial industry.
Candidates have a strong background in dynamic economic modelling. Experience with agent-based or integrated assessment modelling is welcomed, as is knowledge of environmental/energy economics, statistics and patent data analysis.
Using energy to respond to the increased frequency of climate extremes such as heat-waves is a leading example of energy use for adaptation that can lead to broad economic, social, and environmental implications of remarkable scientific and policy relevance. New technologies and behaviours are required in order to ensure that future adaptation needs do not compromise the achievement of mitigation objectives and the transition towards green technologies. Through the integration of energy consumption and patent data with climate science and integrated assessment modelling, the ESR will develop and study models to describe how adaptation-driven energy requirements (behaviours and technologies related to space cooling) will be shaped under a variety of future climate and socioeconomic settings.
(i) map available data sources of energy consumption and patent data related to sustainable cooling, and match them with climate data; (ii) empirical analysis of individual adaptive strategies (diffusion and adoption of sustainable cooling technologies and behaviours); (iii) develop a computational agent-based model integrating big-data on energy use for adaptation (e.g. diffusion and adoption of sustainable cooling technologies and behaviours); (iv) derivation of policy recommendations in relation to the green transition.
The candidate should have a strong background in complex systems science, mathematical economics, statistical physics or a related discipline. Programming skills, affinity with economic dynamic models, network models, game theory and the topic of energy are desirable.
The electricity market is currently in transition from a centralised oligopoly towards a decentralised market with local supply-demand interactions among adaptive agents, including producers, consumers and prosumers. Such markets inherently self-organise. The main research question of this project is: under which market conditions will this self-organisation lead to systemically stable, unstable, or self-organised critical (SOC) dynamics?
Agent-based modelling and the theory of dynamical systems theory will be used to obtain insights into the stability of future electricity markets, under various policy scenarios, while taking into account path dependence. We consider situations where households will produce, store, as well as consume energy locally from each other (`prosumers'), and they are free to ask and bid any price in a local market, auction, or one to one (`over-the-counter'). Eventually this project will help identify the ideal post-transition electricity (target) market, and will provide policy recommendations to facilitate the transition to this situation.
Strong background in dynamic economic modelling (ideally experience with agent-based modelling). Knowledge of behavioral economics and graph/network theory are welcome.
Developing agent-based models (ABMs) with a social network structure to study changes in lifestyles for a low-carbon economy, accounting for bounded rationality and social interactions. This involves making connections with the field of cultural evolutionary analysis. Particular attention will be given to issues such as exemplary conducts, role models, social movements and appropriate framing of information provision. This will result in insights about an effective policy mix, including possibly economic instruments, information provision or nudges, and technical standards.
The developed ABMs will be used to examine if and how fast lifestyles change under certain policy mixes. This involves testing various hypotheses: fostering the status of low-carbon goods drives consumption away from carbon-intensive goods and services; stimulating behavioural diffusion is more effective in reducing emissions than stimulating knowledge diffusion; or, frugality and simpler lifestyles can diffuse through exemplary behaviours.
Strong background in dynamic economic modelling (ideally experience with agent-based modelling). Good knowledge of graph/network theory and interest in the role of information in economic decision making is welcome.
A powerful feature of agent-based models (ABMs) is that they are suitable to combine market-based instruments and information-provision instruments in an integrated model structure, thus allowing to study their potential synergistic effects. This will be done by giving attention to information-provision instruments such as eco-labels, green awards, information campaigns and Internet-based social media, as well as by limiting false or otherwise undesirable information through regulation of advertising of high-carbon goods and services.
The developed models will be used to study distinct policy mixes addressing how market-based instruments propagate through social networks, and what kind of measures (comparative information feedbacks, advertising) are able to magnify their impact. In particular, we will test the hypothesis if it is more effective to stimulate low-carbon alternatives through social marketing strategies or discourage high-carbon ones by regulating commercial advertising.
Strong analytical and computational skills, strong interest in economic modelling, ability to work on both theoretical and empirical problems. Expertise in network analysis and interest in industrial dynamics and innovation are favourably considered.
The aim of this project is to develop models on the long-term evolution of production networks in order to better understand the structural drivers of innovation. We will investigate on the one hand to which extent the direction of technological change can be explained by endogenous industrial dynamics and on the other hand how innovation policy can influence the speed and the direction of technological change.
The project will deliver models useful to develop scenarios and quantitative projections about future technological pathways at the sectoral and/or the micro-economic level. Such models could be used to identify leverage points for industrial policy in production networks and to derive measures of adaptive capacities for economic regions and sectors, notably in the context of the decarbonization of the European economy.
Strong background in dynamic economic modelling and expertise/interest in one or several of the areas of industrial dynamics, evolutionary economics and economics of innovation.
The project examines under which circumstances a coordinated active public investment policy, heavily supporting the development and diffusion of certain (selected) new technologies and approaches, is a necessary and useful approach for fostering technological change and politically desired transition processes. Potential benefits will be compared to the danger of induced lock-in at inefficient technologies in face of complex technological landscapes.
(i) development and analysis of industry-level and macroeconomic computational models with strong microfoundations incorporating processes of technological search by firms and instruments of public intervention, (ii) obtaining insights about the relationship between the complexity of the technological landscape, the information flows between agents and the effectiveness of different types of public technology investments, (iii) derivation of policy recommendations.
The candidate should have a strong background in dynamic economic modelling (also in the area of agent-based simulations) and quantitative economic analysis. Expertise in network analysis and interest in monetary and financial economics are favourably considered.
The project aims at assessing the consequences of financial innovation (especially cryptocurrencies, fintech, new business models for the banking system) for economic stability and the design and the implementation of monetary policy.
(i) development and analysis of the properties of macroeconomic computational models with deep micro-foundations incorporating a key role for a relatively sophisticated financial system (ii) exploration of the effects of “complex” financial institutions/assets/liquidity instruments on economic stability and the effectiveness of different types of monetary policy measures.
The candidate should have good quantitative skills. A background in Quantitative Economics or Econometrics would be optimal, or possibly in Mathematics or Physics with a strong interest in Economics; programming skills and familiarity with macroeconomic, environmental or network models are a plus.
The project examines under which circumstances the development of international collaborations (exchange of knowledge and technology) between developing and developed countries may allow latecomers to engage into a process of “sustainable growth”. In particular, since developing countries will have to engage heavily in the process of emission reductions but the majority of innovations is concentrated in the developed countries, the design of climate policies and the stringency of regulations may foster or hamper the capability of an emerging country to exit the trap of underdevelopment.
(i) Development and exploration of the properties of industry-level and macroeconomic computational models, enriched by network structure of international trade/global value chains. (ii) These models will be used to gain insights about the relationship between the complexity of the technological landscape and the effectiveness of different types of public policies (climate change policies, industrial policies).