This topic invites research on forecasting performance in macroeconomic models, with a particular focus on Dynamic Stochastic General Equilibrium (DSGE) and empirical forecasting frameworks. It is deliberately broad, allowing candidates to define their own research directions within macroeconomic forecasting — including potential work on (Bayesian) prediction averaging and related methodologies.

• The central objective is to compare, combine, and improve forecasts of key macroeconomic variables — such as output growth, inflation, unemployment, and interest rates — across short-, medium-, and long-term horizons. Research may be grounded in a wide range of model classes, including:
• Structural models: DSGE (e.g., medium-scale New Keynesian, DSGE-VAR, or DSGE with time-varying volatility)
• Classical time-series models: ARIMA/AR, VAR
• Bayesian approaches: Shrinkage priors, BVARs with global–local priors, time-varying parameters, and stochastic volatility (TVP-SV)
• Factor-based models: DFMs, FAVAR
• Mixed-frequency and state-space methods: Bridge equations, MIDAS, ragged-edge filtering
• Data-driven and hybrid methods: Machine learning models, as well as survey- or market-based forecasts used as benchmarks or conditioning information

A unifying methodological theme is forecast evaluation under model uncertainty. Possible directions include:

• Transparent forecast comparison (“horse races”): Evaluating real-time and revised data, accounting for publication lags and data vintages.
• Density forecasting and calibration: Assessing probabilistic accuracy using log scores, CRPS, quantile loss, or turning-point metrics.
• Forecast combination and prediction pooling: Developing and testing approaches such as simple forecast averaging, Bayesian Predictive Synthesis, Dynamic Model Averaging, or time-varying combination weights that adapt to evolving regimes.
• Special emphasis may be placed on models’ ability to handle structural change—for example, during inflation surges, supply shocks, or policy regime shifts. Approaches incorporating time variation, regime switching, or judgment-consistent conditioning can be assessed comparatively in this context.

Illustrative (non-exhaustive) research directions:

• Structural vs. empirical performance: Out-of-sample comparisons of DSGE (and DSGE-VAR) forecasts against BVAR, factor, MIDAS, or machine-learning benchmarks, with or without external conditioning (e.g., survey or nowcast inputs).
• Prediction averaging as insurance: Design and evaluation of prediction pools or Bayesian synthesis frameworks that enhance point and density forecast accuracy, and investigation of when sparse versus dense combinations perform best.
• Forecasting under instability: Analysis of whether time-varying parameters, stochastic volatility, or regime-switching mechanisms improve forecast reliability during atypical or turbulent episodes.

The student will have access to high-performance computing resources, enabling the implementation and estimation of computationally intensive models.

The supervisor for this topic is associate professor Jan Čapek. Detailed information about the supervisor, his publications and research projects can be found here. 

This broad topic invites candidates to study how fiscal actions—spending changes, tax reforms, and consolidations—translate into real-economy outcomes, and why those effects differ across time, place, and policy design. The emphasis is empirical and policy-relevant: measuring fiscal multipliers and mapping their state-dependence (business-cycle slack, inflation regime/ZLB, debt levels, openness, exchange-rate regime), composition (spending vs. taxes; investment vs. consumption vs. transfers; PIT/CIT/VAT), and distributional and external effects (employment, inequality, prices, debt sustainability, trade balance). Applicants may focus on the US, the EU/Eurozone, or multi-country panels.

Illustrative research angles (choose one or combine):

• Heterogeneous multipliers: Do GDP, employment, and investment respond differently to the same fiscal shock in recessions vs. expansions, under high inflation or high public debt, or across exchange-rate regimes?
• Composition and design: Which instruments—public investment, targeted transfers, broad tax cuts, corporate tax changes—deliver the largest/most durable effects, and for whom?
• Consolidation strategies: Are expenditure-based vs. tax-based consolidations associated with different short-run costs and long-run debt dynamics?
• Distribution & prices: How do fiscal actions reshape the income distribution, labor market attachment, and inflation dynamics?
• Spillovers & coordination: What are the cross-border and subnational spillovers (e.g., Eurozone, US states), and when does coordination raise multipliers?

Empirical toolbox (students can specialize):

• Causal identification: narrative/event studies (text-as-data from budgets, laws, press releases), external-instrument/“proxy” SVARs, local projections (with interactions/thresholds for nonlinearity), regional quasi-experiments (DiD, shift-share/BBE instruments, grant formulas, base closings), and panel methods with episode-level identification.
• Micro–macro links: household/firm microdata to estimate MPCs and investment responses that aggregate to macro effects; heterogeneity by income, liquidity, age, or sector.
• Expectations and information effects: high-frequency identification around fiscal announcements; sovereign spreads and survey expectations to proxy credibility/fiscal space.

The student will have access to high-performance computing resources, enabling the implementation and estimation of computationally intensive models.

The supervisor for this topic is associate professor Jan Čapek. Detailed information about the supervisor, his publications and research projects can be found here. 

The thesis aims to identify the causes of exchange rate volatility and evaluate its consequences in post-communist countries. The range of exchange rate volatility varied greatly among these countries. One obvious reason is that these countries applied different exchange rate regimes. However, this thesis aims to evaluate the topic more complexly and to identify other determinants. Another task is to consider the impact of this volatility, primarily on the integration of these countries into world markets.

The supervisor for this topic is associate professor Libor Žídek. Detailed information about the supervisor, his publications and research projects can be found here.