ABSTRACT. We contribute to the mixed-frequency regressions literature by introducing an innovative Bayesian approach. Relying on this new methodology, we provide novel empirical evidence of uncertainty shock identification for the US economy. As main findings, we document an ``aggregation bias'' when we adopt a common frequency framework at a monthly frequency instead of estimating a mixed-frequency model with weekly and daily frequencies. The bias is amplified when the COVID-19 pandemic crisis is included.

ABSTRACT. Independent Component Analysis (ICA) is a statistical method that transforms a set of random variables in least dependent linear combinations. Under the assumption that the observed data are mixtures of non-Gaussian and independent processes, ICA is able to recover the underlying components, but a scale and order indeterminacy. Its application to structural vector autoregressive (SVAR) models allows the researcher to recover the impact of independent structural shocks on the observed series from estimated residuals. We analyze different ICA estimators, recently proposed within the field of SVAR identification, and compare their performance in recovering structural coefficients. Moreover, after suggesting an algorithm that solve the ICA indeterminacy problem, we assess the size distortions of the estimators in hypothesis testing. We conduct our analysis by focusing on distributional scenarios that get gradually close the Gaussian case, which is the case where ICA methods fail to recover the independent components. In terms of statistical properties of the ICA estimators, we find no evidence that a method outperforms all others. We finally present an empirical illustration using US data to identify the effects of government spending and tax cuts on economic activity, thus providing an example where ICA techniques can be used for hypothesis testing

ABSTRACT. How large are government spending and tax multipliers? The fiscal proxy-SVAR literature provides heterogenous estimates, depending on which proxies - fiscal or non-fiscal - are used to identify fiscal shocks. We reconcile the existing estimates via a flexible vector autoregressive model that allows to achieve identification in presence of a number of structural shocks larger than that of the available instruments. Our two main findings are the following. First, the estimate of the tax multiplier is sensitive to the assumption of orthogonality between total factor productivity (non-fiscal proxy) and tax shocks. If this correlation is assumed to be zero, the tax multiplier is found to be around one. If such correlation is non-zero, as supported by our empirical evidence, we find a tax multiplier three times as large. Second, we find the spending multiplier to be robustly larger than one across different scenarios characterized by different sets of instruments. Our results are robust to the joint employment of different fiscal and non-fiscal instruments.

ABSTRACT. Despite several attempts to provide a definite pattern regarding the effects of personality traits on performance in higher education, the debate over the nature of the relationship is far from being conclusive. The use of different subject pools and sample sizes, as well as the use of identification strategies that either do not adequately account for selection bias or are unable to establish causality between measures of academic performance and noncognitive skills, are possible sources of heterogeneity. This paper investigates the impact of the Big Five traits, as measured before the beginning of the academic year, on the grade point average achieved in the first year after the enrolment, taking advantage of a unique and large dataset from a cohort of Italian students in all undergraduate programs containing detailed information on student and parental characteristics. Relying on a robust strategy to credibly satisfy the conditional independence assumption, we find that higher levels of conscientiousness and openness to experience positively affect student score.

ABSTRACT. We assess the effect of financial and in-kind aid programs on the location decision process of students. This phenomenon is analyzed by using a unique dataset with administrative data on Italian university students enrolled for the first time in the academic year 2014-2015 along with detailed information on the need-based policies comprised by the Diritto allo studio universitari program. We explicitly consider the heterogeneity in students' preferences concerning these policies. In particular, we take advantage of the Latent Class Logit approach to model systematic and random heterogeneity in preferences conditioned on students' individual characteristics. The estimated parameters are exploited to quantify the sensitivity of student's location choices by computing willingness to pay and semi-elasticity measures. Our results suggest that policies that provide scholarships and places in dormitories together positively affect students' choice probabilities, this way suggesting that these policies are effective in attracting more students to specific universities. Semi-elasticities results indicate that students with better high school diploma grades are more sensitive to these policies.

ABSTRACT. Neighborhoods can have a considerable impact on children's future outcomes, but the mechanisms through which they operate are not well understood yet. I use public housing regenerations as a natural experiment that changes the composition of more deprived neighborhoods. In London, many public housing buildings have been demolished in the past two decades to pave the way for new developments with higher housing density, giving home to about 160,000 people. These regenerations caused little displacement for the students living in the buildings slated for demolition, and the new developments were targeted mostly by more affluent households. I have constructed a novel database by geocoding all regenerations in Greater London and linking them to administrative records on primary school-age students. I compare the achievement of students in schools of similar neighborhoods but located at different distances from the regeneration before and after its completion, and use a grandfathering instrument to estimate the impact on students who were already in a school close to a regeneration before its completion ('incumbents'). I find that incumbent students exhibit higher test scores in math and language at the end of primary school after the regeneration, with stronger gains for more disadvantaged and low-ability students. The empirical evidence suggests that such gains are driven by changes in the demand for schools due to the inflow of more affluent parents with strong preferences for school quality.

ABSTRACT. Performance standards in need-based student aid may exacerbate inequality in higher education, but at the same time they might improve efficiency of aid expenditure, for instance by increasing on-time graduation. Disentangling the effect of the two building blocks of student aid (“need” and “merit”) is therefore of key interest to policy makers, yet it is difficult to achieve since aid comes as a complex package. In this paper, we seek to estimate the causal effect of tightening the merit-based requirements of need-based student aid on short-term and long-term student academic performance. This is done leveraging a reform in an Italian region that increased by 40% (i.e. from 25 to 35 out of a maximum of 60) the number of credits to be passed in the first academic year to maintain aid eligibility. Using administrative data from an Italian university, this study reveals that tightening merit requirements had a statistically significant, positive effect on various dimensions of performance of the “average” student in aid. Aided students in cohorts after the reform acquire more formative credits in the first year, along with better grades. Moreover, these effects persist over time and are conducive to a higher probability to obtain the BA degree on time. However, an analysis of treatment heterogeneity unveils winners and losers from the policy: the positive effects are indeed concentrated among high ability students, while our findings point to a decrease in re-enrolment of low-ability students in student aid. The results have clear policy implications, corroborating the idea that carefully balancing need and merit is a pressing issue for the design of student aid packages that preserve both equity and efficiency of the university system.

ABSTRACT. This paper separately identifies financial and uncertainty shocks using a novel SVAR procedure and discusses their distinct monetary policy implications. The procedure relies on the qualitatively different responses of corporate cash holdings: after a financial shock, firms draw down their cash reserves as they lose access to external finance, while uncertainty shocks drive up cash holdings for precautionary reasons. Although both financial and uncertainty shocks are contractionary, my results show that the former are inflationary while the latter generate deflation. I rationalize this pattern in a New-Keynesian model: after a financial shock, firms increase prices to raise current liquidity; after an uncertainty shock, firms cut prices in response to falling demand. These distinct channels have stark monetary policy implications: conditional on uncertainty shocks the divine coincidence applies, while in case of financial shocks the central bank can close the output gap at the cost of more unstable inflation.

ABSTRACT. This paper quantifies the effects of equity tail risk on the US government bond market. We estimate equity tail risk as the option-implied stock market volatility that stems from large negative jumps as in Bollerslev, Todorov and Xu (2015), and assess its value in reduced-form predictive regressions for Treasury returns and an affine term structure model for interest rates. We find that the left tail volatility of the stock market significantly predicts one-month excess returns on Treasuries both in- and out-of-sample. The incremental value of employing equity tail risk as a return forecasting factor can be of economic importance for a mean-variance investor trading bonds. The estimated term structure model shows that equity tail risk is priced in the US government bond market. Consistent with the theory of flight-to-safety, we find that (i) Treasury prices increase and (ii) funds flow from equities into bonds when the perception of tail risk is higher. Our results concerning the predictive power and pricing of equity tail risk extend to major government bond markets in Europe.

ABSTRACT. We explore empirically the transmission of U.S. financial and macroeconomic uncertainty to emerging market economies (EMEs). We start by assuming that there are crucial differences between volatility and uncertainty, and between the latter and its shocks. With the help of Bayesian vector autoregressions and narrative sign restrictions, we first identify two measures of U.S. uncertainty shocks, which appear to explain the dynamics of output developments better than conventional volatility measures. Next, we find evidence that adverse shocks to U.S. aggregate uncertainty are associated with marked contractions in most EMEs’ business cycles. However, we detect significant cross-country heterogeneity in the responses to U.S uncertainty shocks. We also find generalized declines in stock market values, which supports the so-called Global Financial Cycle hypothesis.

ABSTRACT. This paper quantifies the finance uncertainty multiplier (i.e., the magnifying effect of the real impact of uncertainty shocks due to financial frictions) by re- lying on two historical events related to the US economy, i.e., the large jump in financial uncertainty occurred in October 1987 (which was not accompanied by a deterioration of the credit supply conditions), and the comparable jump in finan- cial uncertainty in September 2008 (which went hand-in-hand with an increase in financial stress). Working with a VAR framework and a set-identification strategy that focuses on - but it is not limited to - restrictions related to these two dates, we estimate the finance uncertainty multiplier to be equal to 2, i.e., credit supply disruptions are found to double the negative output response to an uncertainty shock. We then employ our model to estimate the overall economic cost of the COVID-19-induced uncertainty shock under different scenarios. Our results point to the possibility of a cumulative yearly loss of industrial production as large as 31% if credit supply gets disrupted. Liquidity interventions that keep credit con- ditions as healthy as they were before the COVID-19 uncertainty shock are found to substantially reduce such loss.

ABSTRACT. We propose an individual level health index that quantifies the latent evolving health stock of individuals in the Danish population. We construct the index using five detailed population-wide registers for individuals of ages 50 to 80. In the methodology, we rely on simple as well as more advanced techniques such as Charlson's comorbidity index, principal components analysis, regularized logistic regression, random forests, feed forward neural networks and stacked autoencoders. The indices are dissected through importance rankings to understand driving mechanisms used to establish the health scores. To assess the usefulness of our health indices, we evaluate their ability to create scores that allow for an accurate stratification of individuals in terms of their observed risk of relevant health outcomes. Then, we use these indices to forecast mortality risk over different horizons and test for their predictive accuracy. Our analysis indicates that general healthcare indicators processed through nonlinear methods determine individual health states well. Finally, we show that our health indices lead to favorable out-of-sample results in individual-level mortality forecasting.

ABSTRACT. This work investigates whether and to what extent the amalgamation of local health departments may affect population health using the most recent available Italian data on mortality (2003-2017) at the municipality level. We set up a quasi-experimental research design using an event study framework that enables us to test for potential pre-existing trends as well as phase-in policy impact. Because Italian regions implemented the consolidation reform in different years, our estimates rely on a staggered two-way fixed effect approach. We deal with the issue of treatment effect heterogeneity over time and across municipalities by adopting an interaction-weighted estimator. Potential mechanisms are unfolded by considering mortality rates in different age groups and for six leading causes of death. We find that population mortality has increased in the years following the policy change, suggesting that merging in the health care sector might have unintended remarkable effects on health.

ABSTRACT. Response management to the SARS-CoV-2 outbreak requires to answer several forecasting tasks. For hospital managers, a major one is to anticipate the likely needs of beds in intensive care in a given catchment area one or two weeks ahead, starting as early as possible in the evolution of the epidemic. This paper proposes to use a bivariate Error Correction model to forecast the needs of beds in intensive care, jointly with the number of patients hospitalised with Covid-19 symptoms. Error Correction models are found to provide reliable forecasts that are tailored to the local characteristics both of epidemic dynamics and of hospital practice for various regions in Europe in Italy, France and Scotland, both at the onset and at later stages of the spread of the disease. This reasonable forecast performance suggests that the present approach may be useful also beyond the set of analysed regions.

ABSTRACT. We compare the short- to medium-term forecasting accuracy of mortality rates and life expectancy of six differents methods. These include the simple random walk (benchmark), five well-established methods (Lee-Carter, Lee-Miller, Booth-Maindonald-Smith, Hyndman-Ullah) and a general factor model. The comparison is carried out on three populations with different characteristics (France, Italy and USA) using data for 1950 onwards for estimation and up to 2016 for evaluation under both symmetric and asymmetric loss functions. We also carry out a counterfactual experiment designed to reproduce the likely impact of Covid-19 pandemic on forecasts using 2020 as a jump-off year. Our main conclusion is that no model can be singled out as clearly superior to the others, but broadly speaking the performances of the Lee-Miller variant of the Lee-Carter model and the Hyndman and Ullah model are often better than those of the other models.

ABSTRACT. We consider an extension of ARCH (∞) models to account for conditional asymmetry in the presence of high persistence. After stating existence and stationarity conditions, this paper develops the statistical inference of such models and proves the consistency and asymptotic distribution of a Quasi Maximum Likelihood estimator. Some particular specifications are studied and tests for asymmetry and GARCH validity are derived. Finally, we present an application on a set of equity indices to reexamine the preeminence of GARCH (1, 1) specifications. We find strong evidences that the short memory feature of such models is not suitable for lightly traded assets.

ABSTRACT. In this paper we develop a time-varying extension of the Poisson autoregressive model with exogenous covariates (TV-PARX), originally introduced by Fokianos et al. (2009) and Agosto et al. (2016). We show that the score-driven framework is particularly suitable to recover the evolution of time-varying parameters and provides the required flexibility to model and forecast time series of counts characterised by convoluted nonlinear dynamics and structural breaks. We study the asymptotic properties of the TV-PARX model and prove that, provided some mild conditions, maximum likelihood estimation yields strongly consistent and asymptotically normal parameter estimates. Finite-sample performance and forecasting accuracy are evaluated through Monte Carlo simulations. The empirical usefulness of the time-varying specification of the TV-PARX model is shown by analysing and forecasting the US corporate default counts.

ABSTRACT. In many real-world applications, time series of counts are commonly observed given the discrete nature of the variables of interest. This paper introduces a new stochastic process with values in the set Z of integers with sign. The increments of process are Generalized Poisson differences and the dynamics has an autoregressive structure. We study the properties of the integer-valued GARCH process introduced, and exploit the thinning representation to derive stationarity conditions, the stationary distribution of the process and its conditional and unconditional moments. We provide a Bayesian inference framework and an efficient posterior approximation procedure based on Markov Chain Monte Carlo. Numerical illustrations on simulated data show the effectiveness of the proposed inference. The applications to accidents data and cyber threats data show that the proposed model is well suited for capturing persistence in the conditional moments and in the over-dispersion feature of the data.

ABSTRACT. In this paper we develop novel bootstrap methods for point processes modeled through their dynamic intensity function, as in the case of Hawkes process. We introduce a new, fast and easy to implement, `fixed intensity' bootstrap (FIB), where the intensity function of the bootstrap samples matches the intensity function of the original data, and hence is fixed across bootstrap replications. We address the issue of robustness of the bootstrap by also proposing and discussing non-parametric bootstrap schemes which directly resample a time-change transformation of the original event times, rather than drawing them from the exponential distribution, as in parametric algorithms.

ABSTRACT. This paper uses a new textual data index for predicting stock market data. The index is applied to a large set of news to evaluate the importance of one or more general economic related keywords appearing in the text. The index assesses the importance of the economic related keywords, based on their frequency of use and semantic network position. We apply it to the Italian press and construct indices to predict Italian stock and bond market returns and volatilities in a recent sample period, including the COVID-19 crisis. The evidence shows that the index captures well the different phases of financial time series. Moreover, results indicate strong evidence of predictability for bond market data, both returns and volatilities, short and long maturities, and stock market volatility.

ABSTRACT. The uncertainty surrounding economic forecasts is generally related to multiple sources of risks, of domestic and foreign origin. This paper studies the predictive distribution of Italian GDP growth as a function of selected risk indicators, related to both financial and real economic developments. The conditional distribution is characterized by means of expectile regressions. Expectiles are closely related to the Expected Shortfall, a well-known measure of risk with desirable properties. Here a decomposition of Expected Shortfall in terms of contributions of different indicators is provided, which allows to track over time the main drivers of risk. Our analysis confirms that financial conditions are relevant for the left tail of the predictive distribution but it also highlights that indicators of global trade and uncertainty have strong explanatory power for both left and right tail. Overall, our findings suggest that Italian GDP risks have been mostly driven by foreign developments around the Great Recession, by domestic financial conditions at the time of the Sovereign Debt Crisis and by economic policy uncertainty in more recent years.

ABSTRACT. This paper proposes a novel asymmetric continuous probabilistic score (AS) for evaluating and comparing density forecasts. It extends the proposed score and defines a weighted version, which emphasizes regions of interest, such as the tails or the center of a variable's range. A test is also introduced to statistically compare the predictive ability of different forecasts. The AS, is of general use in any situation where the decision maker has asymmetric preferences in the evaluation of the forecasts. In an artificial experiment, the implications of varying the level of asymmetry in the AS, are illustrated. Then, the proposed score and test are applied to assess and compare density forecasts of macroeconomic relevant datasets (US employment growth) and of commodity prices (oil and electricity prices) with particular focus on the recent COVID-19 crisis period.

ABSTRACT. This paper proposes tests for pairwise and multiple out-of-sample comparisons of parametric conditional quantile models when the goal is to forecast intervals such as in the Growth-at-Risk (GaR) literature. We evaluate (possibly misspecified) models in terms of the conditional coverage w.r.t. the union of information sets of forecast intervals. The test is implemented by ranking the distance between actual and nominal coverage across models for some loss. Our approach operates uniformly over a compact set of quantile ranks, acknowledging that models' relative forecast ability may vary over different quantile subsets. We establish weak convergence to a zero mean Gaussian process uniformly over that set of quantile ranks, and prove the first order validity of block bootstrap critical values. The latter represents a viable alternative over more standard subsampling procedures which are not appropriate for many shorter macroeconomic time series. We apply our test to shed further light on the recent GaR literature. We find that the standard linear quantile regression approach for quarterly Gross Domestic Product (GDP) growth performs no better than a competing threshold quantile model, and is uniformly dominated over lower quantiles when extending the analysis to study the timelier monthly GaR of industrial production.

ABSTRACT. Expanding parental choice in education may increase system-wide productivity if parents select schools that are a specifically good match for their children. I investigate this hypothesis by studying the effect of attending the school of choice on student achievement in London primary schools. I exploit as good as random variation in parental preference for school arising from centralised assignment which, in case of oversubscription, awards school offer by residential distance. I replicate the algorithm used for assignment and compare students around year-specific catchment boundaries that cannot be exactly anticipated by parents. I find that attending the school of choice increases student achievement compared to an institution with lower parental preference but similar value-added. Results suggest that parents pick schools that are specifically effective in increasing their children’s achievement, improving the efficiency of school seats allocation. One potential explanation of match effects I uncover is that parents of low-ability males select schools with lower peer quality and these better suit their learning needs.

ABSTRACT. By exploiting rich retrospective data on childhood immunization, socioeconomics, and health status in China (the China Health and Retirement Longitudinal Study), we assess the long-term effects of childhood vaccination on cognitive and educational outcomes in that country. To do so, we apply various techniques (e.g., propensity score and coarsened exact matching and correlated random effects) to different sets of conditioning variables and subsamples to estimate the average treatment on the treated effect of childhood vaccination. Our results confirm that vaccinations before the age of 15 have long-term positive and economically meaningful effects on nonhealth outcomes such as education and cognitive skills. These effects are relatively strong, with vaccinated individuals enjoying about one more year of schooling and performing substantially better later in life on several cognitive tests.

ABSTRACT. The measure of the Education quality is fundamental to provide information to the public opinion and to give the policymakers the opportunity to implement policies aimed to improve the education supply and the teaching and learning process. To find models that provide a valid measure of the effectiveness of school quality and of Education Value Added (VA) is crucial for the Accountability. The aim of this paper is to compute and evaluate a VA measure for Italian Primary school and compare the performance of different econometric approaches. Moreover, considering the usefulness of VA measures in evaluate the school efficiency, generating an incentive to cheating or teaching the test and then affecting its results, the role of monitoring is also investigated. The study will use the INVALSI standardized tests in Italian for the cohort of students at grade II (Primary) in 2011-12; 2012-13 and the cohort of students at grade V-(Primary) in 2014-15 2015-16. The reached results make in evidence that higher test scores are driven by the test scores in the previous period and the parents’ educational level also play a role. According to the expectations, a significative negative sign coefficient is estimated for the presence of monitoring. The School Value Added distribution have also derived for different areas.

ABSTRACT. The large literature on the determinants of students' performance have recently stressed the importance of the school's leadership and organization. In this paper we aim to assess the impact of a series of reforms of the school institutional setting in Italy that resulted in a change from a full-time school principal in each school to a sort of part-time one, with single school principal managing different schools located in dierent areas. This cost-saving policy may have negatively aected education outcomes. To test this hypothesis, we exploit the program eligibility rule that states that schools with a number of students enrolled below a certain threshold have to merge with others, i.e. share the same school principal, and apply the Fuzzy Regression Discontinuity Design approach with a continuity-based local polynomial approximation to identify the effect on the students' performance. We focus on dropouts and grade retention rates as outcomes of interest. Estimation results show negative eects in Italian upper secondary schools, especially for treated schools far away from the directional school, i.e. where the school principal is based.

ABSTRACT. There exists a significant differential in life satisfaction between disabled and nondisabled people, to the disadvantage of the former. The present work considers both satisfaction and meaning of life (as different facets of happiness), investigating whether environmental accessibility mediates the relationship between disability and happiness. Furthermore, the effect of accessibility on the happiness of different categories of disabled is analysed. The environmental accessibility index is built using data from the 2012 Eurobameter survey on accessibility, while the rest of the variables come from the EU-SILC 2013, which includes an ad hoc module on well-being. Findings show that higher environmental accessibility narrows the happiness gap between disabled people and the rest, even after interaction terms between disability and economic status are introduced. Moreover, environmental accessibility has a greater impact on the happiness of older disabled people, while the opposite is true of disabled people in the highest income quartile.

ABSTRACT. Tertiary education is a major source of human capital accumulation. As people become increasingly mobile, students become more willing to cross the borders. However, the literature on this phenomenon’s economic impact is quite scant. Moreover, internationally mobile students in tertiary education represent a source of international human capital, which is increasingly demanded by firms investing overseas. This paper investigates the relationship between mobility of international students in tertiary education and foreign direct investment (FDI) using bilateral data on mergers and acquisitions (M&A) deals and international students’ flows among G20 countries. The results suggest a positive relation between international students and bilateral cross-border M&A activity. This relation holds true when measuring the bilateral cross-border M&A activities using either the number or total value of deals. Our findings indicate that the impact of international students on M&A deals depends on the development level of both the acquirer and target countries.

ABSTRACT. Following in the footsteps of the literature on "empirical welfare maximization", this paper wants to contribute by stressing the policymaker perspective via a practical illustration of an optimal policy assignment problem. More specifically, by focusing on the class of "threshold-based" policies, we first set up the theoretical underpinnings of the policymaker selection problem, to then offer a practical solution to this problem via an empirical illustration using the popular LaLonde (1986) training program dataset. The paper proposes a pseudo-code implementation of the solution procedure that can be easily programmed with standard statistical software.

ABSTRACT. In this paper, we estimate the spatial spillovers mechanism across 20 Italian regions using the default rates on loans facilities as proxy of the loans probability of default, over the period 1996-2015. The data, at quarterly frequency, are available for consumer households, non-financial firms and producer households. First, we investigate the presence of spatial dependence across the regional loan default rates. Second, we evaluate whether the Mezzogiorno regions are more affected by spillover effects arising from the Northern regions. For this purpose, we use the connectedness measures proposed by Diebold & Yilmaz (2012) and by Greenwood-Nimmo et al. (2015), which are based on the generalized forecast error variance decomposition (GFEVD) obtained from the estimation of a Vector Autoregression model. Given the relatively large number of variables, we use the Adaptive elastic net to estimate the VAR model. The empirical findings reveal an increase in default rates spatial dependence over the 2011Q4 - 2015Q4 (crisis) period, especially for producer households. Moreover, we find evidence of a strong dependence of the Islands from the North of Italy, while the other Southern regions are found to be the most contributor, together with the Northwest of Italy, of financial distress to the remaining macro-regions.