ABSTRACT. Using a small Bayesian dynamic factor model of the euro area we estimate the deviations of output from its trend that are consistent with the behavior of inflation. We label these deviations the output gap. In order to pin-down the features of the model, we evaluate the accuracy of real-time inflation forecasts from different model specifications. The version that forecasts inflation best implies that after the 2011 sovereign debt crisis the output gap in the euro area has been much larger than the official estimates. Versions featuring a secular-stagnation-like slowdown in trend growth, and hence a small output gap after 2011, do not adequately capture the inflation developments.

ABSTRACT. Jin, Corradi and Swanson (2016) develop a forecast evaluation methodology which is robust to the choice of the loss function and they establish a map between loss function robust forecast evaluation and stochastic dominance. However, their tests are not uniformly valid and have correct asymptotic size only under the least favorable case under the null. Since, tests for stochastic dominance can be seen as tests for infinitely many moment inequalities, we use tools from Andrews and Shi (2013, 2016) to develop test for robust forecast comparison which are uniformly asymptotically valid and asymptotically non-conservative. The (Many) Moment Inequalities literature mainly focus on independent case. In our set-up, forecast error are typically non martingale difference sequence, because of dymanic misspecification. We establish uniform convergence (over error support) of HAC estimators and of their bootstrap counterpart. Futhermore, we extend Generalized Moment Selection tests for the presence of non-vanishing recursive and rolling estimation error. The suggested testing procedure is used to evaluate SPF, and in particular to assess whether there are subset of forecasters which produce more accurate prediction, in a loss free manner.

ABSTRACT. This paper proposes a new method of calculation the whole day covariance matrix that includes information from the overnight period. Proposed estimator is based on the scaling-and-weighting estimator of Hansen&Lunde 2005, extended to the multivariate case. For the scaling, a new concept of matrix proportionality is proposed. Three ways of optimal weighting of matrices are introduced, each of them exploit the theorem from the paper of Hansen&Lunde 2005. New method decreases the noise of the whole day covariance estimator in comparison existing approach: a sum of intraday realized covariance and outer product of overnight returns. The performance of the two estimators is compared using both simulated and real data.

* Hansen, Lunde 2005 A realized variance for the whole day based on intermittent high-frequency data.

ABSTRACT. From two professional news providers we retrieve news stories and earnings announcements of the S&P 100 constituents and a set of 10 macroeconomic fundamentals. We thus create an extensive and innovative dataset which contains information with minute precision, useful to analyze the link netween news and asset price dynamics. We develop a novel text-analysis technique to detect the sentiment of a financial text of any type, size and audience, and propose a set of more than 4K news-based variables that provide natural proxies of the information used by heterogeneous market players. We first shed light on the impact of news on daily realized volatility and select news measures by penalized regression. Then, we distinguish the relative importance between news measures and use them to forecast volatility.

ABSTRACT. The aim of this paper is to quantify the strength and the direction of volatility spillovers between five EMU stock markets over the 2000-2016 period. For this purpose we apply the Diebold and Yilmaz (2012) methodology, based on a generalized forecast error variance decomposition, to downside and upside realised semi-volatility series. While the analysis of Diebold and Yilmaz (2012) is based on a stationary VAR, we take into account the long-memory behaviour of the series,by using the multivariate extension of the HAR model (named VHAR model). Finally, we show how the choice of the normalization rule can bias the spillover computation in a full sample as well as in a rolling sample analysis.

ABSTRACT. The weighted-average least squares (WALS) approach, introduced by Magnus et al. (2010) in the context of Gaussian linear models, has been shown to enjoy important advantages over other strictly Bayesian and strictly frequentist model averaging estimators when accounting for problems of uncertainty in the choice of the explanatory variables. In this paper we extend the WALS approach to deal with the uncertainty surrounding the specification of the linear predictor in the wider class of generalized linear models (GLMs). We study the large-sample properties of the WALS estimator for GLMs under a local misspecification framework that allows the development of asymptotic model averaging theory. Moreover, we investigate the finite sample properties of this estimator by a Monte Carlo experiment whose design is based on the real empirical analysis of attrition in the first two waves of the Survey of Health, Ageing and Retirement in Europe (SHARE).

ABSTRACT. We study the semi-parametric estimation of the conditional mode of a random vector that has a continuous conditional joint density with a well-defined global mode. A novel full-system estimator is proposed and its asymptotic properties are studied allowing for possibly dependent data. We specifically consider the estimation of vector autoregressive conditional mode models and of structural systems of linear simultaneous equations defined by mode restrictions. The proposed estimator is easy to implement and simulation results suggest that it is well behaved in finite samples. A simple empirical example is provided to illustrate the application of the proposed methods.

ABSTRACT. We characterise the mean and variance parameters that distributionally misspecified maximum likelihood estimators can consistently estimate in multivariate conditionally heteroskedastic dynamic regression models. We also provide simple closed-form consistent estimators for the rest. The inclusion of means and the explicit coverage of multivariate models make our procedures useful not only for GARCH models but also in many empirically relevant macro and finance applications involving VARS and multivariate regressions. We study the statistical properties of our proposed consistent estimators, as well as their efficiency relative to Gaussian pseudo maximum likelihood procedures. Finally, we provide finite sample results through Monte Carlo simulations.

ABSTRACT. Most literature about the causes of revolts and civil conflicts has been focused on the role of the lack of freedom and the greater instability generated by a heterogeneous society. However, the uprising of a revolution seems to suffer largely from collective action problems. This may explain why people privately support a revolt, but are unwilling to act unless they observe others doing so, and why grievance rarely leads to revolutions. Following MacCulloch and Pezzini (2010), we focus on the micro-structure of revolutionary propensity, taking into account the extent of country fractionalization and a measure of autonomy freedom. Two are our main research questions: first, is a higher degree of societal heterogeneity correlated with larger grievance? Second, does an autonomous behavior mitigate individual’s proclivity towards revolutionary processes? We answer these research questions allowing for differences in religious affiliation, socio-economic characteristics and the structure of the society.

ABSTRACT. In this paper we examine the extent to which college decisions among adolescents depend on the decisions of their peers. In particular, we ask whether individuals derive utility from conformity in college enrollment and academic mobility. We propose a new methodology in mitigating reflection and endogeneity issues in identifying social interactions. We use the proportion of females in a student's last year's reference group (school and neighbourhood) as an instrumental variable. We investigate utility spillovers from the educational choices of students in consecutive cohorts. Spatial variation allows us to identify social interactions in groups of various sizes. We use a new dataset that spans the universe of high school graduates. We find positive and significant externalities in the decision to enrol in college and the decision to migrate to a different city among peers that belong to the same social group. Results indicate that students who are in a school or neighborhood with 100% more peers who enrolled in college last year are 29% or 9.6% more likely to themselves attend college.

ABSTRACT. This paper estimates the causal impact of police on crime. To solve the problem of reverse causality, I use the natural experiment represented by the creation of a special policing unit to intensify surveillance around a few tournament-related locations within the city of São Paulo during the 2014 FIFA World Cup. The tournament did not impact crime only by means of a policing change: higher fan concentration likely boosted criminal activity, while the voluntary incapacitation of a substantial number of individuals engaged in watching the games expectedly acted in the opposite direction. Exploiting the fact that the World Cup affected different areas of the city through different channels and at different times, I manage to isolate the impact of interest. Difference-in-differences estimates reveal that policing has a large and significant deterrent effect on crime.

ABSTRACT. We investigate then impact of submarket concentration in the decision of launching a "new" product in the US pharmaceutical 3 digit submarkets during the period 1987-1998. We introduce additional regressors inside the entry-exit standard reduced form models such as the concentration, the company relative size compared to incumbent firms, the number of competing products, all obtained by exploiting the submarkets information. The results, obtained using a logit model, are based on an annual panel whit two cross-sectional dimensions: companies and submarkets, and they are directed to a role of the submarket concentration as barrier to entry.

ABSTRACT. Strict exogeneity of covariates other than the lagged dependent variable, conditional on unobserved heterogeneity, is often required for consistent estimation of binary panel data models. However, this assumption is likely to be violated in practice because of feedback effects from the past of the outcome variable on the present value of covariates and no general solution is yet available. We introduce a likelihood ratio test for strict exogeneity, and for its modification that accommodates noncausality, in binary panel data with fixed T derived from the definitions provided by Chamberlain (1982) for nonlinear models. Our approach builds on a novel Pseudo-Conditional Maximum Likelihood estimator based on an approximation of the multiple-lag dynamic logit model. We further show, by simulation, that a specialization of the approximating model to a first-order Markov process for the dependent variable provides a consistent estimator of the state dependence parameter, especially in short panels with a large number of cross-section units.

ABSTRACT. This paper explores the possibility of using axioms of microeconomic theory to improve the Discrete Choice Experiments design and facilitate inference. Transitivity of preferences is a necessary assumption for analyzing DCE-collected data in a Random Utility Theory framework. Monotonicity is often used when interpreting data and it seems a safe assumption in a wide range of applications. Then, it would make sense to use these assumptions at the design stage, with the objective of avoiding inefficient questions (choice sets). However, violations of these axioms are found in literature. In this paper we first assess the reasons This paper assesses their relevance through an experiment and it discusses how current RUT models deal with the issue.

ABSTRACT. In this paper we study how monetary policy, economic uncertainty and economic policy uncertainty impact on the dynamics of gross capital inflows in the US. Particular attention is paid on the mixed frequency-nature of the economic time series involved in the analysis. A MIDAS-SVAR model is presented and estimated over the period 1988-2013. While no relation is found when using standard quarterly data, exploiting the variability present in the series within the quarter shows that the effect of a monetary policy shock is greater the longer the time lag between the month of the shock and the end of the quarter. In general, the effect of economic and policy uncertainty on US capital inflows are negative and significant. Finally, the effect of the three shocks is different when distinguishing between financial and bank capital inflows from one side, and FDI from the other.

ABSTRACT. We develop a news-media textual measure of aggregate economic uncertainty – the fraction of Financial Times articles containing uncertainty-related keyphrases – for 1982–2014 at daily to annual frequencies. We contribute to the literature in three areas.

First, we provide a measurement framework that links observed expressions of uncertainty in newspaper articles to a latent propensity to express uncertainty, which we argue is an ordinal proxy for the uncertainty that matters for economic decision-making, namely the intensity of the cognitive state of uncertainty. We use this framework to estimate how the noise-to-signal ratios varies with sample size (or frequency) and show that noise variance is modest at monthly and lower frequencies, and approaching signal variance at daily frequency.

Second, we study key choices in the empirical implementation of such measures more deeply than has been done previously, focusing on uncertainty keyphrase selection, isolating economic uncertainty, de-duplication of articles, and appropriate scaling of the uncertainty measure, with a critique of scaling methods commonly used in the literature. Our findings provide empirical foundations for the extant literature, and evidence-based recommendations for methodological improvements.

Third, we conduct the first detailed comparative analysis of a news-media uncertainty measure with another uncertainty proxy, stock returns volatility. Our narrative analysis establishes the plausibility of our news-media measure. Our quantitative analysis reveals a strong relationship to stock volatility on average. But this relationship breaks down periodically, with timing that suggests that the semantics of the word “uncertainty” may be biased towards downside uncertainty or risk. Finally, we establish the absence of Granger causation between the measures down to daily frequency, except for a one-day lead of stock volatility over news-media uncertainty, which is to be expected given that the FT is published before the market opens.

ABSTRACT. We propose a Bayesian panel model for mixed frequency data whose parameters can change over time according to a Markov process. Our model allows for both structural instability and random effects. We develop a proper Markov Chain Monte Carlo algorithm for sampling from the joint posterior distribution of the model parameters and test its properties in simulation experiments. We use the model to study the effects of macroeconomic uncertainty and financial uncertainty on a set of variables in a multi-country context including the US, several European countries and Japan. We find that for most of the variables financial uncertainty dominates macroeconomic uncertainty. Furthermore, we show that uncertainty coefficients differ if the economy is in a contraction regime or in an expansion regime.

ABSTRACT. We estimate marginal propensities to consume from wealth shocks for Italian households in the early part of the Great Recession. Large asset price shocks in 2008 underpin an IV estimator. A euro fall in risky financial wealth resulted in cuts in annual total (non-durable) consumption of 8.5-9 (5.5-5.7) cents. There is evidence of effects on food spending. Responses of total and non-durable spending to changes in housing wealth are 0.2 to 0.3 cents/euro. Point estimates of the effect of the financial wealth shock are larger if the youngest and/or oldest households are excluded. Results indicate that responses to the wealth shock were stronger for those who became pessimistic about the stock market, and for those owners of risky assets who also held mortgage debt. Counterfactuals indicate financial wealth effects were important (relative to other factors) for consumption falls in Italy in 2007/08.

ABSTRACT. Can political contributions buy influence? This paper studies whether firms trade political contributions for public procurement contracts. To answer this question, I focus on the Lithuanian political economy. Combining data on a large number of government tenders, the universe of corporate donors and firm characteristics, I examine how a ban on corporate donations affects the awarding of procurement contracts to companies that donated in the past. Consistent with political favoritism, contributing firms' probability of winning goes down by five percentage points as compared to that of non-donor firms after the ban. Among different mechanisms, the hypothesis that corporate donors get confidential information on competing bids prevails. The empirical results are in line with predictions from a first-price sealed-bid auction model with one informed bidder. Evidence on firm bidding and victory margins suggests that contributing firms adjust their bids in order to secure contracts at a maximum revenue. I assess that tax payers save almost one percent of GDP thanks to the reform.

ABSTRACT. According to "agenda setting theory" increased media coverage of an issue makes people believe that the issue is of particular importance. In this paper we test this hypothesis by analyzing the impact of corruption salience in the media on individuals' perceptions about the extent of corruption. To this purpose, we exploit information on individuals' perceptions of the likelihood that corruption events may occur in everyday life and combine it with a dataset containing the number of news items related to corruption that appeared on the homepage of the websites of the 30 most widely read national and local newspapers on the day on which the individual was interviewed. Results show that a one standard deviation increase in the number of corruption news items reported by the newspapers causes an increase in corruption perceptions of about 3.5 per cent and a decrease in trust in justice effectiveness of about 5.2 per cent. We provide evidence that these effects are mainly driven by persuasion rather than learning in that perceptions respond more to news unrelated to facts than to those which report arrests, investigations or convictions. We finally suggest that increasing transparency in the institutions may be more effective in reducing the bias in corruption perceptions generated by the media, than increasing competition in the local media market.

ABSTRACT. Combining a Danish panel of yearly administrative wealth reports with the unexpected timing of sudden parental deaths, I exploit inheritance episodes to characterize wealth accumulation dynamics in the ten years following a financial windfall. Consistent with the predictions of a buffer stock model of consumption, liquid assets quickly converge to pre-inheritance levels. However, real estate and financial investments persist over time. Age and liquidity constraints do not explain these results: Heirs exploit inheritance to accumulate housing equity if young, and precautionary savings if liquidity constrained. These causal estimates highlight the importance of consumption models capable of distinguishing between multiple assets.

ABSTRACT. We propose a RESET-type test for the null hypothesis of linearity of a cointegrating relationship with an asymptotic chi-squared null distribution. The test is based on an extension of the Integrated Modified OLS estimator of Vogelsang and Wagner (2014) from linear cointegrating relationships to multivariate cointegrating polynomial relationships. For the case of full design we furthermore provide fixed-b asymptotic theory for our RESET test. The theoretical results are complemented by a small simulation study.

ABSTRACT. We investigate the asymptotic and finite sample properties of the most widely used information criteria for co-integration rank determination in 'partial' systems, i.e. in co-integrated Vector Autoregressive (VAR) models where a sub-set of variables of interest is modeled conditional on another sub-set of variables. The asymptotic properties of the Akaike Information Criterion (AIC), the Bayesian Information Criterion (BIC) and the Hannan-Quinn Information Criterion (HQC) are established, and consistency of BIC and HQC is proved. Notably, consistency of BIC and HQC is robust to violations of the hypothesis of weak exogeneity of the conditioning variables with respect to the co-integration parameters. More precisely, BIC and HQC recover the true co-integration rank from the partial system analysis also when the conditional model does not convey all information about the co-integration parameters. This result opens up interesting possibilities for practitioners who can now determine the co-integration rank in partial systems without being concerned with the weak exogeneity of the conditioning variables. A Monte Carlo experiment which considers large systems as data generating process shows that BIC and HQC applied in partial systems perform reasonably well in small samples and comparatively better than 'traditional' methods for co-integration rank determination. We further show the usefulness of our approach and the benefits of the conditional system analysis to co-integration rank determination with two empirical illustrations, both based on the estimation of VAR systems on U.S. quarterly data. Overall, our analysis clearly shows that the gains of combining information criteria with partial systems analysis are indisputable.

ABSTRACT. This paper presents generalizations of Granger's Representation Theorem for I(1) processes and of Johansen's Representation Theorem for I(2) systems for processes integrated of any (possibly fractional) order. General inversion results are given, which gives explicit expressions for the coefficients of the vector Equilibrium Correction representation in terms of their Common Trends representation, and vice-versa. General representation theorems are then proved on the relationship between the inversion, the local Smith form and the Jordan structure. This allows a unified treatment of different representation results in the literature, with explicit expressions of the matrix coefficients.