Search

Inspired by the classical riot model proposed by Granovetter in 1978, we consider a parametric stochastic dynamical system that describes the collective behavior of a large population of interacting agents. By controlling a parameter, a policy maker seeks to minimize her own disutility, which in turn depends on the steady state of the system. We show that this economically sensible optimization is ill-posed and illustrate a novel way to tackle this practical and formal issue. Our approach is based on theΓ-convergence of a sequence of mean-regularized instances of the original problem. The corresponding minimum points converge toward a unique value that intuitively is the solution of the original ill-posed problem. Notably, to the best of our knowledge, this is one of the first applications ofΓ-convergence in economics.

This paper attempts to fill the gap between economic and behavioral literature on tax evasion. We propose a micro-founded model whose results are consistent with the ‘Slippery Slope’ Framework and with the Expected Utility Theory. Individual taxpayers face the problem of deciding the fraction of income they wish to disclose to the government, based on their trust and perceived power of authorities. Trust is modeled as the voluntary compliance originated by the umbrella-term tax morale and power is shaped by the perceived enforcement, mainly motivated by individual risk-aversion. Furthermore, we made use of Agent-Based simulations to replicate the ‘slippery slope’ conditions which allowed us to test the effects of different parameters on tax evasion. Compliance is primarily enhanced by tax morale and risk aversion, while it is secondarily motivated by higher audit probabilities and penalty fees. Tax rates, however, play a negative effect on tax compliance, as agents are less willing to pay taxes whenever facing larger obligations. Additionally, we study taxpayer behavior when the audit rate is zero and when agents are inclined to make charitable donations. Above all, we derive the conditions under which individuals would fully-evade, partially evade, fully-comply or even over-comply as charitable giving in the absence of audits.

We present a verification, an extension and a reanalysis of “Uncertain outcomes and climate change policy”, R. Pindyck, Journal of Environmental Economics and Management, 2012. As far as verification is concerned, we are able to reproduce the results provided in Pindyck’s work in many cases and convincingly confirm the quality of the work. Some discrepancies are present, they are due to rounding or related to specific sets of parametric values and do not change the economic interpretation or significance of the results. The re-estimation of the model with more recent data on climate change made available in 2014 shows that temperature increments are now deemed to be higher in mean but less dispersed. As a consequence, the willingness to pay doesn’t vary much with respect to the original paper. We also modify the functional form describing the impact of temperature increase on the growth rate of consumption and obtain much bigger and potentially problematic increments of the willingness to pay. Finally, the paper demonstrates that the numerical results are sensitive to a variety of technical settings used in the computations and suggests that great care is needed in obtaining estimates and employing results in policy discussions.

In this paper we investigate the effects of using a Facebook page exclusively devoted to a first year Mathematics course in a large Italian public university. Posts and discussions supported traditional face-to-face lectures and students could freely post queries and get help from professors and peers. We use a newly constructed dataset to measure how this influences the grade they achieved and the probability of getting a passing grade. Firstly, we find that active students, who read and post more often, perform slightly better than non active ones, who mainly read the content, but the effect is not significant. However, other measures of activity, such as the frequency of visualization of the page significantly increase the probability of earning a passing grade, after controlling for students� characteristics and robust ex-post measures of ability. Secondly, we exploit a quasi-natural experiment to compare the performance of students having access to Facebook with that of a large sample of similar students who were not offered the support page in another branch of the uni- versity. Difference-in-difference estimates show that students who could access online discussions gain on average 1 additional point out of 30. The effect is, hence, significant but rather small and of possibly limited practical relevance.

A trader's execution strategy has a large effect on his profits. Identifying an optimal strategy, however, is often frustrated by the complexity of market microstructure's. We analyse an order book based continuous double auction market under two different models of trader's behaviour. In the first case actions only depend on a linear combination of the best bid and ask. In the second model traders adopt the Markov perfect equilibrium strategies of the trading game. Both models are analytically intractable and so optimal strategies are identified by the use of numerical techniques. Using the Markov model we show that, beyond the best quotes, additional information has little effect on either the behaviour of traders or the dynamics of the market. The remarkable similarity of the results obtained by the linear model indicates that the optimal strategy may be reasonably approximated by a linear function. We conclude that whilst the order book market and strategy space of traders are potentially very large and complex, optimal strategies may be relatively simple and based on a minimal information set.

We model a continuous double auction with heterogenous agents and compute approximate optimal trading strategies using evolution strategies. Agents privately know their values and costs and have a limited time to transact. We focus on equilibrium strategies that are developed taking into account the number of traders that submitted orders previously, as well as the number of who will submit subsequently. We find that it is optimal to place increasingly aggressive orders, according to a roughly linear schedule, and test the resulting equilibrium for robustness and accuracy.