How manipulable are prediction markets?
Speaker: Roberto Rozzi (University of Siena)
Date: April 15, 2025
Time: 2:00 PM - 3:30 PM
Location: Centrale Aula 5, Scuola Superiore Sant'Anna
For online participation click here
To book your slot to meet the speaker, click here
Abstract: In this paper, we conduct a large-scale field experiment to investigate the manipulability of prediction markets. The main experiment involves randomly shocking prices across 817 separate markets; we then collect hourly price data to examine whether the effects of these shocks persist over time. We find that prediction markets can be manipulated: the effects of our trades are visible even 60 days after they have occurred. However, as predicted by our model, the effects of the manipulations somewhat fade over time. Markets with more traders, greater trading volume, and an external source of probability estimates are harder to manipulate.
Founding conditions, learning and the economic success of young firms
Speaker: Massimo Riccaboni (IMT Lucca)
Date: April 8, 2025
Time: 2:00 PM - 3:30 PM
Location: Maffi Aula 14, Scuola Superiore Sant'Anna
For online participation click here
To book your slot to meet the speaker, click here
Abstract: It is well-known that the probability of survival increases with firm age and that long-lived firms are more productive than short-lived ones. We advance two theories of the economic success of new firms – noisy learning and capability learning - to predict and explain the economic success of young new firms. To do so, we use high-dimensional administrative data for the whole population of newly incorporated firms in the Netherlands from 2006-2021. We apply machine learning techniques in the framework proposed by Mueller and Spinnewijn (2023) to quantify the heterogeneity of firm success probability types and interpret results with theories of firm learning. Our results suggest that young firms’ success probabilities are highly persistent, both over the firm’s life course and over the business cycle. We conclude that dynamic selection significantly influences the gradient between economic success and age, confirming the predictions of noisy learning theory. We find no evidence of a relationship between changes in firm success probabilities and early life conditions, refuting the predictions of capability learning theory. Overall, our results strongly support noisy learning as the primary mechanism of young firm learning.
What drives the recent surge in inflation? The historical decomposition roller coaster
Speaker: Nicolo Maffei-Faccioli (Norges Bank)
Date: April 3, 2025
Time: 4:30 PM - 6:00 PM
Location: Boyl aula Talento all'Opera, Scuola Superiore Sant'Anna
For online participation click here
Abstract: What drives the recent inflation surge? To answer this question, one must decompose inflation fluctuations into the contribution of structural shocks. We document how whimsical such a historical shock decomposition can be in standard vector autoregressive (VAR) models. We show that the deterministic component of the VAR tends to be imprecisely estimated, making the shock contributions poorly identified under general conditions. Our preferred approach to solve this problem— the single-unit-root prior—can massively shrink the uncertainty around the estimated deterministic component. Once this uncertainty is taken care of, demand shocks unambiguously appear as the main drivers of the inflation surge in the United States, the euro Area, and in four small open economies.
Peer effect analysis with latent processes
Speaker: Vincent Starck (University of Munich)
Date: March 25, 2025
Time: 2:00 PM - 3:30 PM
Location: Boyl aula 2, Scuola Superiore Sant'Anna
For online participation click here
Abstract: I analyze peer effects arising from irreversible decisions in the absence of a standard social equilibrium. I model a latent sequence of decisions in continuous time and obtain a closed-form solution for the likelihood. The method avoids regression on conditional expectations or linear-in-means regression -- and thus reflection-type problems (Manski, 1993) or simultaneity issues -- by modeling the (unobserved) realized direction of causality, whose probability is identified. For implementation, I propose a parsimonious parametric specification that introduces a peer effect parameter meant to capture the causal influence of first-movers on their peers. Parameters are shown to be consistently estimated by maximum likelihood methods and lend themselves to standard inference under repeated network asymptotics.