Working papers results

This paper investigates the differential response of male and female voters to competitive persuasion in political campaigns. During the 2011 municipal elections in Milan, a sample of eligible voters was randomly divided into three groups. Two were exposed to the same incumbent's campaign but to different opponent's campaigns, with either a positive or a negative tone. The third-control-group received no electoral information. The campaigns were administered online and consisted of a bundle of advertising tools (videos, texts, slogans). Stark gender differences emerge. Negative advertising increases men's turnout, but has no effect on women. Females, however, vote more for the opponent and less for the incumbent when they are exposed to the opponent's positive campaign. Exactly the opposite occurs for males. Additional tests show that our results are not driven by gender identification with the candidate, ideology, or other voter's observable attributes. Effective strategies of persuasive communication should thus take gender into account. Our results may also help to reconcile the conflicting evidence on the effect of negative vs. positive advertising, as the average impact may wash out when aggregated across gender.

Rational voters update their subjective beliefs about candidates' attributes with the arrival of information, and subsequently base their votes on these beliefs. Information accrual is, however, endogenous to voters' types and difficult to identify in observational studies. In a large scale randomized trial conducted during an actual mayoral campaign in Italy, we expose different areas of the polity to controlled informational treatments about the valence and ideology of the incumbent through verifiable informative messages sent by the incumbent reelection campaign. Our treatments affect both actual vote shares at the precinct level and vote declarations at the individual level. We explicitly investigate the process of belief updating by comparing the elicited priors and posteriors of voters, finding heterogeneous responses to information. Based on the elicited beliefs, we are able to structurally assess the relative weights voters place upon a candidate's valence and ideology. We find that both valence and ideological messages affect the first and second moments of the belief distribution, but only campaigning on valence brings more votes to the incumbent. With respect to ideology, cross-learning occurs, as voters who receive information about the incumbent also update their beliefs about the opponent. Finally, we illustrate how to perform counterfactual campaigns based upon the structural model.

In a decision problem under uncertainty, a decision maker considers a set of alternative actions whose consequences depend on uncertain factors outside his control. Following Luce and Raiffa (1957), we adopt a natural representation of such situation that takes as primitives a set of conceivable actions A, a set of states S and a consequence function from actions and states to consequences in C. With this, each action induces a map from states to consequences, or Savage act, and each mixed action induces a map from states to probability distributions over consequences, or Anscombe-Aumann act. Under a consequentialist axiom, preferences over pure or mixed actions yield corresponding preferences over the induced acts. The most common approach to the theory of choice under uncertainty takes instead as primitive a preference relation over the set of all Anscombe-Aumann acts (functions from states to distributions over consequences). This allows to apply powerful convex analysis techniques, as in the seminal work of Schmeidler (1989) and the vast descending literature. This paper shows that we can maintain the mathematical convenience of the Anscombe-Aumann framework within a description of decision problems which is closer to applications and experiments. We argue that our framework is more expressive, it allows to be explicit and parsimonious about the assumed richness of the set of conceivable actions, and to directly capture preference for randomization as an expression of uncertainty aversion.

We examine a number of unexplored factors that affect the ex-post adoption rates of newly listed stock options. We show that a variety of measures of information asymmetries for underlying stocks predict option adoption rates. This occurs even when we control for factors that have been found to be significant in earlier literature, such as stock volatility and volume. However, option listings induce a reduction in the strength of the information asymmetries in the underlying stock. Further, option bid-ask spreads start from low initial levels and increase over time, which is consistent with a modest initial aggressiveness of informed investors.

We examine asset allocation decisions under smooth ambiguity aversion when an investor has a prior degree of belief in an asset pricing model (e.g., the domestic CAPM). Different from the Bayesian portfolio approach, in our model the investor separately relies on the conditional distribution of returns and on the posterior over uncertain parameters to make asset allocation decisions, rather than on the predictive distribution of returns that integrates priors and likelihood information in a single distribution. This is a key feature implied by smooth ambiguity preferences. We find that in the perspective of US investors, ambiguity aversion can generate strong home bias in their equity holdings, regardless of their belief in the domestic CAPM or of their degree of risk aversion. Our results extend and become stronger under regime-switching investment opportunities.

We investigate the lead-lag relationships between issuer- and investor-paid credit rating agencies, in the aftermath of the regulatory reforms undertaken in the U.S. between 2002 and 2006 -including watch list inclusions and outlooks. First, we find that the lead effect of investor-paid over issuer-paid credit rating agencies has weakened: in recent years, causality has turned bi-directional. Second, when changes in outlooks are included, we find evidence of a less conservative behavior by issuer-paid agencies, when compared to their rating behavior. Third, stock prices manifest statistically significant abnormal reactions to downgrades of all agencies; however, abnormal negative returns are significantly higher for investor-paid downgrades. Our results support the hypothesis that when issuer-paid agencies have seen their market power threatened by tighter regulations, they have felt incentives to improve the quality and timeliness of their ratings. However, event studies show that markets still price stocks under the assumption that investor-paid rating actions carry superior information.

This paper employs a recent statistical algorithm (CRAGGING) in order to build an early warning model for banking crises in emerging markets. We perturb our data set many times and create "artificial" samples from which we estimated our model, so that, by construction, it is flexible enough to be applied to new data for out-of-sample prediction. We find that, out of a large number (540) of candidate explanatory variables, from macroeconomic to balance sheet indicators of the countries' financial sector, we can accurately predict banking crises by just a handful of variables. Using data over the period from 1980 to 2010, the model identifies two basic types of banking crises in emerging markets: a "Latin American type", resulting from the combination of a (past) credit boom, a flight from domestic assets, and high levels of interest rates on deposits; and an "Asian type", which is characterized by an investment boom financed by banks' foreign debt. We compare our model to other models obtained using more traditional techniques, a Stepwise Logit, a Classification Tree, and an "Average" model, and we find that our model strongly dominates the others in terms of out-of-sample predictive power.

In the theory of psychological games it is assumed that players' preferences on material consequences depend on endogenous beliefs. Most of the applications of this theoretical framework assume that the psychological utility functions representing such preferences are common knowledge. But this is often unrealistic. In particular, it cannot be true in experimental games where players are subjects drawn at random from a population. Therefore an incomplete-information methodology is called for. We take a first step in this direction, focusing on models of guilt aversion in the Trust Game. We consider two alternative modeling assumptions: (i) guilt aversion depends on the role played in the game, because only the 'trustee' can feel guilt for letting the co-player down, (ii) guilt aversion is independent of the role played in the game. We show how the set of Bayesian equilibria changes as the upper bound on guilt sensitivity varies, and we compare this with the complete-information case. Our analysis illustrates the incomplete-information approach to psychological games and can help organize experimental results in the Trust Game.

Using personal data collected on the internet, firms and political campaigners are able to tailor their communication to the preferences and orientations of individual consumers and voters, a practice known as hypertargeting. This paper models hypertargeting as selective disclosure of information to an audience with limited attention. We characterize the private incentives and the welfare impact of hypertargeting depending on the wariness of the audience, on the intensity of competition, and on the feasibility of price discrimination. We show that policy intervention that bans the collection of personally identifiable data (for example, through stricter privacy laws requiring user consent) is beneficial when consumers are naive, competition is limited, and firms are able to price discriminate. Otherwise, privacy regulation often backfires.

Fiscal consolidations achieved by means of spending cuts are much less costly in terms of output losses than tax-based ones. The difference cannot be explained by accompanying policies, including monetary policy, and it is mainly due to the different response of business confidence and private investment. We obtain these results by studying the effects of the adoption of fiscal consolidation plans (rather than isolated shocks), that is combinations of tax increases and spending cuts, some unanticipated, other anticipated, in a sample of 16 OECD economies.