I am an assistant professor of economics at Corvinus University of Budapest.
I am interested in the problem of (persistent) disagreement. My theoretical approach is subjective models and misspecified learning. I build on the social psychology literature about naive realism and the theory of construal. (earlier interests: production networks and COVID economics)
CV, SSRN, Google Scholar, oliver.racz[a]uni-corvinus.hu
Challenging the Ruler by Rallying the Voter: Evidence from Hungary Electoral Studies (June 2026)
Joint with Tamás Kovalcsik and Mátyás Bódi [Online Appendix] [Replication Material] [BibTeX]
We quantify the causal effect of opposition rallies (Peter Magyar's Country Tour) under autocratic media control in Hungary on the 2024 EP elctions. We find that 11 percent of Fidesz (Viktor Orban) voters turned away in response to Péter Magyar's Country Tour. Find out more...
Economic Costs of Distancing Policy Interventions, Economics and Human Biology (Dec 2025)
[Preprint] [Data and Codes] [BibTeX]
I estimate the short-run economic effects of COVID distancing interventions accounting for their direct (DPI-compliant) and indirect (DPI-triggered voluntary) effects. I find significant output losses but negligible inflationary or unemployment effects. Results suggest the costless adjustment of the composition but not the stringency of the policy mix. Find out more...
Using confidence indicators for the assessment of the cyclical position of the economy MNB Bulletin, (June 2012)
[BibTeX]
Compliance and Containment
Joint with László Tőkés (Corvinus University of Budapest)
Distancing policy interventions are the primary containment tool during pandemics before pharmaceutical solutions become available. This paper addresses a key challenge in the evaluation of non-pharmaceutical interventions: the confounding of policy-compliant and voluntary behavioral responses to NPIs. We develop an empirical strategy that separates these two channels and estimates their containment effects independently. Identification exploits the discontinuous decline in social activity following the introduction of first-wave COVID-19 interventions. Using daily data from 100+ countries, we recover policy-compliant and voluntary components of distancing and estimate their effects on the effective reproduction number. Results reveal substantial heterogeneity in the epidemiological effectiveness of distancing. Voluntary and policy-compliant responses account for broadly similar shares of the observed decline in social activity, yet policy-compliant distancing is about twice as effective per unit of social activity reduction in lowering the reproduction number. Distancing policies targeting specific locations, implemented early and under stronger institutional environments, generate substantially larger reductions in virus reproduction. The contribution of stringent measures was found to be marginal relative to recommendations. Our results suggest that the epidemiological effectiveness of distancing depends not only on its magnitude but also on how, when, and under what institutional conditions it is generated.
Drivers of Link Formation in Production Networks
Joint with Lajos Tamás Szabó (Central Bank of Hungary)
What makes a supplier more attractive than its competitors? Which firms are better at holding on to their buyers or suppliers? Such questions determine the shape of the production network of an economy by driving the formation and separation of input-output links. This paper aims to fill a gap in the literature by providing novel descriptive facts about the drivers of link formation using firm-to-firm transaction data from Hungary between 2014 and 2022. We compare firms engaging in link formation and separation to similar firms not involved in such linking decisions. We find that relatively better-performing (profitable, productive, faster growing) firms are more successful in attracting new and keeping existing buyers and suppliers. Contracting a new buyer is preceded by a surge in investments and new hires, suggesting a well-prepared decision. Firms losing a buyer crank up hiring (sacrificing efficiency) in the last year before they lose their buyer. The evolution of the production network is driven by firm growth, productivity, and efficiency, enhancing aggregate productivity and network concentration.
Speed of Propagation in Production Networks
(Earlier versions available at SSRN: https://ssrn.com/abstract=4392276 or http://dx.doi.org/10.2139/ssrn.4392276 )
The Dot-com bubble, the Global Financial Crisis, and the post-COVID inflation were heavily influenced if not entirely triggered by localized sector-specific shocks, such as a financial market collapse, or the supply disruptions of important commodities. These aggregate events eventually unfolded as the triggering shocks' effects propagated through production networks. This paper aims to quantify the propagation speed by developing a novel econometric method. Estimations are carried out on an annual sample of 66 US industries between 1998 and 2020. The speed of propagation is identified from the propagation of industry TFP shocks within the production network into industry output and prices. Results show that the propagation time – the time shocks need to get transmitted between industries – was between 7 to 8 months on average and it accelerated after the GFC. These results provide evidence for dynamic shock propagation, which production network models have mostly overlooked.
Income Taxation, Transfers and Labour Supply at the Extensive Margin.
Joint with Péter Benczúr, Gábor Kátay and Áron Kiss.
Banque de France Working paper series 487; 2014. p. 1-27. JRC89930 [BibTeX]
This paper estimates the effect of income taxation and transfers on labour supply at the extensive margin, i.e., the labour force participation. We extend existing structural form methodologies by considering the effect of both taxes and transfers. Non-labour income contains the (hypothetical) transfer amount someone gets when out of work, while the wage is replaced by the difference between net wages and the amount of lost transfers due to taking up a job (gains to work). To incorporate these components of the budget set, we employ a detailed tax-benefit model. Using data from the Hungarian Household Budget Survey (HKF), we find that participation probabilities are strongly influenced by transfers and the gains to work, particularly for low-skill groups and the elderly. Moreover, the same change in the net wage leads to a much larger change in the gains to work for low earners, making them even more responsive to wages and taxation. Overall, we find that a single equation can capture a large heterogeneity of individual responsiveness to taxes and transfers. Our parametric estimates can be readily utilized in welfare evaluations, or microsimulation analyses of tax and transfer reforms.
Polarization by Naive Realism
Subjective Models and Naive Realism
joint with Miklós Pintér