ABSTRACT. The need for understanding the propagation mechanisms behind the recent nancial crises lead the increased interest for works associated with systemic risks. In this framework, network-based methods have been used to infer from data the linkages between institutions (or companies. Part of the literature postulates that systemic risk is strictly related (if not equal) to systematic risk. In this work, we elaborate on this hypothesis and introduce a modelling framework where systemic and systematic risks co-exist. The model is a variation of the traditional CAPM/APT model where networks are used to infer the exogenous and contemporaneous links across assets. The systematic risk component acts in an additive way on both the systematic and idiosyncratic risk components. Our proposed methodology is veried both on simulations as well as on real data.

ABSTRACT. Pricing kernel specications reflecting market preferences is used to adjust the risk neutral density of the stock return dynamics. This yields the density of the returns under the physical measure, which captures market expectations. The novelty of the method is that it allows for the non-monotonicity of the pricing kernel which might lead to improved forecasts of the physical density. Based on the DAX Index and option data at EUREX we investigate the existence of the empirical pricing kernel paradox, i.e. a locally increasing pricing kernel, and estimate the physical density of the DAX Index returns. Further on, we investigate the forecasting performance of the estimated probabilities. The results show that our method provides better forecasts of the physical densities as compared to methods using strictly decreasing pricing kernel specication.

ABSTRACT. This paper tests for herd behavior towards the market consensus for a wide set of international stock markets. We adopt all the three currently available approaches to test for the presence of herding behavior towards the market consensus across global stock markets. We find that herding is present only in some emerging markets when we consider more standard models. As soon as we filter out the effects of both a global and a local volatility factor we find that herd behavior is present in almost all stock markets in our data set and that herding is more pronounced during up markets rather than during down markets. We find that herding is a time-varying phenomenon and we analyze the same issue using both the Hwang and Salmon's (2004) approach and a Markov-Regime-Switching model where we allow for a high and a low volatility regime. We also find that herding is more pronounced during low volatility regimes.

ABSTRACT. In this paper we use the Diebold Yilmaz (2009 and 2012) methodology to construct an index of connectedness among five European stock markets: France, Germany, UK, Switzerland and the Netherlands, by using volatility risk premia. The volatility risk premium is measured by the difference between the implied volatility and expected realized volatility of the stock market for next month. While Diebold and Yilmaz focus is on the forecast error variance decomposition of stock returns or range based volatilities employing a stationary VAR in levels, we account for the long memory stationary properties of the levels of volatility risk premia series. Therefore, we estimate and invert a Fractionally Integrated VAR model to compute the cross forecast error variance shares necessary to obtain the index of total connectedness and the net contribution of each series to total connectedness.

ABSTRACT. This paper provides a methodology based on General Variance Decomposition and Bayesian Model Averaging to estimate the degree of economic interconnectedness across different regions, and applies such methodology to a sample of 199 European NUTS2 regions in the period 1980-2008. The estimated connectedness appears very heterogeneous and not symmetric. The idiosyncratic component is not very significant, as well as the common component. A clear pattern of core-periphery exists but not defined in geographical terms. The country component is not very significant, very heterogeneous across countries, and proportional to countries' size. The degree of interconnectedness positively depends on the time horizon of the analysis. Finally, the comparison of the estimated connectedness matrix with two spatial matrices generally used in spatial econometrics (a first-order contiguity and a distance-based matrix) reveals that both are far from representing the actual interconnectedness between European regions.

ABSTRACT. Estimation of knowledge spillovers is a research topic that presents several challenges. Usually, the strategy to identify these spillovers relies on pooling R&D by technological groups, using geographical or technological distance among firms, or using geographical variation in R&D investment laws as an instrument. These strategies present some weaknesses in the presence of correlated unobserved shocks. In this paper we propose a way to estimate knowledge spillovers using excess variance analysis. We model the R&D investment decision as a game of strategic substitutes. The covariance of the best response function, when conditioned on the set of neighboring firms' R&D investment, allows to disentangle between genuine spillovers and correlated shocks. The combination of all the variances and covariances, together with equation in levels, allows to estimate the strength of the spillovers and the importance of the correlated shocks using GMM methods.

ABSTRACT. We propose a novel nonparametricmethodology to estimate how cross-country differences in investment rates, stocks of human capital, growth rates of employment and initial levels of productivity explain the change in the distribution of labour productivity for a sample of 84 countries in the period 1960-2008. We find that the initial level of labour productivity and the investment rate have decreased dispersion and polarization (to large extent only the first); the stock of human capital and growth rate of employment have increased dispersion and polarization (only marginally the first); and that unobservable characteristics of countries have had a modest impact on dispersion, but have played a crucial role in the emergence of polarization in 2008.

ABSTRACT. Increasing globalization and interconnection among countries generates spatial and temporal dependence which will affect the production process of each country. Many studies have analyzed the effect of cross-sectional dependence by using restrictive parametric models. We use a flexible nonparametric two-step approach on conditional efficiencies to eliminate the dependence of production inputs/outputs on these common factors. By using a dataset of 44 countries over 1970-2007, we estimate the global frontier and explore the channels under which Foreign Direct Investment (FDI) and time affect the production process and its components: impact on the attainable production set (input-output space), and the impact on the distribution of efficiencies. We extend existing methodological tools - flexible non parametric location-scale frontier model - to examine these interrelationships. We emphasize the usefulness of ``pre-whitened'' inputs/outputs to obtain more reliable measure of productivity and efficiency to better investigate the driven forces behind the catching-up productivity process. Furthermore, since the influence of external factors has been eliminated, our proposed approach mitigates the problem of endogeneity bias caused by reverse causality between the external factors as FDI and productivity.

ABSTRACT. This paper discusses identification of systems of simultaneous cointegrating equations with integrated variables of order two or higher. Rank and order conditions for identification are provided for general linear restrictions, as well as for equation-by-equation constraints. The conditions are illustrated on models of aggregate consumption with liquid assets and on system of equations for inventories.

ABSTRACT. We review the I(2) model with a focus on its application to near-I(2) data, i.e. I(1) data with a second root very close to unity, and report the results of some Monte Carlo experiments. We show that with I(2) data tests on the long-run coefficients in the I(2) model have small sample properties consistent with asymptotic results. More importantly, we show also that with near-I(2) data the properties of the tests are (i) similar to those found with genuine I(2) data, (ii) systematically superior to those of the analogous tests constructed in the I(1) model, even if the latter is in principle correctly specified and the former is not. Our results thus provide strong support to the suggestion to model near-I(2) data using the I(2) model.

ABSTRACT. In this paper we investigate the asymptotic and finite sample properties of a number of methods for estimating the co-integration rank in integrated vector autoregressive systems of unknown autoregressive order driven by heteroskedastic shocks. We allow for both conditional and unconditional heteroskedasticity of a very general form. We establish the conditions required on the penalty functions such that standard information criterion-based methods, such as the Bayesian information criterion [BIC], when employed either sequentially or jointly, can be used to consistently estimate both the co-integration rank and the autoregressive lag order. We also extend the corpus of available large sample theory for the conventional sequential approach of Johansen (1995) and the associated wild bootstrap implementation thereof of Cavaliere, Rahbek and Taylor (2014) to the case where the lag order is unknown. In particular, we show that these methods remain valid under heteroskedasticity and an unknown lag length provided the lag length is first chosen by a consistent method, again such as the BIC. The relative finite sample properties of the different methods discussed are investigated in a Monte Carlo simulation study. For the simulations DGPs considered, we find that the two best performing methods are a wild bootstrap implementation of the Johansen (1995) procedure implemented with BIC selection of the lag length and an approach which uses the BIC to jointly select the lag order and the co-integration rank. An empirical illustration is given using the term structure of interest rates in the US.

ABSTRACT. Elliot (1998) made an investigation of the test for γ = γ0 based on cointegration techniques when in fact yt in not I(1) but near I(1) in the sense of Phillips (1987). He pointed out that the size or rejection probability could be very sensitive to the parameter which measures the nearness to I(1), and the correlation between error and regressor. We want in this note to supplement with some simulations which shows that if we put the test for γ = γ0 in the context of model based inference, the size distributions are not quite as severe as when just applying the cointegration methodology by asking if y2t − γ0y1t is a stationary long-run relation. Thus we “investigate the χ2 inference on long-run relationships when we are not sure the variables have a unit root”, see Elliott (1998).

THE PAPER IS STILL PRELIMINARY

ABSTRACT. The cross-section average (CA) augmentation approach of Pesaran (A Simple Panel Unit Root Test in Presence of Cross-Section Dependence, Journal of Applied Econometrics 22, 265–312, 2007) and Pesaran et al. (Panel Unit Root Test in the Presence of a Multifactor Error Structure, Journal of Econometrics 175, 94–115, 2013), and the PANIC approach of Bai and Ng (A PANIC Attack on Unit Roots and Cointegration, Econometrica 72, 1127–1177, 2004) and Bai and Ng (Panel Unit Root Tests with Cross-Section Dependence: A Further Investigation, Econometric Theory 26, 1088–1114, 2010) are among the most popular “second-generation” approaches for cross-section correlated panels. One feature of these approaches is that they have different strengths and weaknesses. The purpose of the current paper is to develop PANICCA, a combined approach that exploits the strengths of both CA and PANIC.

ABSTRACT. Recently, an investigation of unobserved and time-varying multilateral resistance and omitted trade determinants has assumed a prominent role in order to measure the Euro effects on trades precisely. We implement two methodologies: the factor-based gravity model by Serlenga and Shin (2013) and the spatial-based techniques by Behrens, Ertur and Kock (2012), both of which allow trade flows and error terms to be cross-sectionally correlated. Applying these approaches to the dataset over 1960-2008 for 190 country-pairs of 14 EU and 6 non-EU OECD countries, we find that the Euro impact estimated by the factor-based model amounts to 4-5% only, far less than 20% estimated by the spatial-based model. The cross-section dependency test results also confirm that the factor-based model is more appropriate in accommodating correlation between regressors, and unobserved individual and time effects. Overall we may conclude that the trade-creating effects of the Euro should be viewed in the proper historical and multilateral perspective rather than in terms of the formation of a monetary union as an isolated event.

ABSTRACT. We propose a test for state dependence in binary panel data with individual covariates. For this aim, we rely on a quadratic exponential model in which the association between the response variables is accounted for in a different way withrespect to more standard formulations. The level of association is measured by a single parameter that may be estimated by a conditional maximum likelihood approach. Under the dynamic logit model, the conditional estimator of this parameter converges to zero when the hypothesis of absence of state dependence is true. This allows us to implement a t-test for this hypothesis which may be very simply performed and attains the nominal significance level under any structure of the individual covariates.

Through an extensive simulation study, we find that our test has good finite sample properties and it is more robust to the presence of (autocorrelated) covariates in the model specification in comparison with other existing testing procedures for state dependence. The proposed test is illustrated by two empirical applications: the first is based on data coming from the Panel Study of Income Dynamics and concerns employment and fertility; the second one is based on the Health and Retirement Study and concerns the self reported health status.

ABSTRACT. Within the framework of dynamic panel data models with mean stationarity, one additional moment condition may remarkably increase the efficiency of the system GMM estimator. This additional condition is essentially a condition of "homoskesdasticity" of the individual effects; it is "implicitly satisfied" in all the Monte Carlo simulations on dynamic panel data models available in the literature (including the experiments with heteroskedasticity, which is always confined to the idiosyncratic errors), but not "explicitly" exploited. Monte Carlo experiments show remarkable efficiency improvements when the distribution of individual effects, and thus of the initial observation yi0, are skewed, thus including the very important cases in economic applications that include variables like individual wages, sizes of the firms, number of employees, etc.

ABSTRACT. We introduce a multivariate estimator of Financial volatility that is based on the theory of Markov chains. The Markov chain framework takes advantage of the discreteness of high-frequency returns, and we show that the estimator is consistent with a Gaussian limit distribution. Furthermore, the Markov chain framework provides simple expres- sions for both the estimator and its asymptotic variance. We study the Önite sample properties of the estimation in and extensive simulation study, and the estimator is applied to high-frequency data from the period that the global Financial crisis unfolded.

ABSTRACT. The Multiplicative Error Model (Engle (2002)) for nonnegative valued processes is specified as the product of a (conditionally autoregressive) scale factor and an innovation process with nonnegative support. A multivariate extension allows the innovations to be contemporaneously correlated. The estimation procedure is hindered by the lack of sufficiently flexible probability density functions for such processes. We adopt copula functions to be able to estimate the parameters of the scale factors and of the correlations of the innovation processes. An application on the interaction between volatility, volume and number of trades illustrates the procedure.

ABSTRACT. We provide clear-cut evidence for economically and statistically significant multivariate jumps (multi-jumps) occurring simultaneously in stock prices by using a novel nonparametric test based on smoothed estimators of integrated variances. Detecting multi-jumps in a panel of liquid stocks is more statistically powerful and economically informative than the detection of univariate jumps in the market index. On the contrary of index jumps, multi-jumps can indeed be associated with sudden and large increases of the variance risk-premium, and possess a statistically significant forecasting power for future volatility and correlations which implies a sizable deterioration in the diversification potential of asset allocation.

ABSTRACT. This paper introduces a new modelling for detecting the presence of commonalities in a set realized volatility time series. In particular, we propose a multivariate generalization of the heterogeneous autoregressive model that is endowed with a common index structure. Our Vector Heterogeneous Autoregressive Index model (or VHARI) can be simply estimated by a proper switching algorithm when applied to realized volatility or bi-power variation measures of the latent volatility feature. We illustrate its usefulness when modelling the realized volatilities of 20 international equity indexes.

ABSTRACT. This paper identifies the role of filling station density for choosing green cars, in order to disentangle the chicken-and-egg problem that arises when developing new technologies, i.e. the market share of alternative cars will be likely to increase as the density of alternative filling stations increases, and viceversa. I provide an empirical framework to analyze the role of filling station location in car demand and the entry decision by a filling station. I combine them to compare the effect of an incentive on the fuel supply side (reducing sunk costs to new entrants only for the alternative fuels offered) with the subsidies on the price of alternative fuel cars. First results suggest a fundamental role of filling station density in cars adoption.

ABSTRACT. This paper studies the impact of local property taxation on firms performance: employment, sales and total factor productivity. In line with Duranton et al. (2011), we propose a spatial transformation of the data to largely reduce the endogeneity of local taxation. Our approach also controls for unobserved time-invariant establishment heterogeneity and unobserved time-varying site-specific effects. We apply our methodology to a panel data of Italian manufacturing firms in 2001-2010. Moreover (and differently from Duranton et al., 2011) we employ a new set of instruments based on the political alignment of each specific jurisdiction with the national government. Our results show that local taxation has a negative impact on firm performances with an elasticity of about 0.7. We test the presence of agglomeration rents a la Baldwin-Krugman finding that the effect of taxation is much lower in denser jurisdictions. We perform several robustness check in order to rule out confounding effects as the presence of multi-plants, the possibility to relocate or the co-existence between property taxes and other business taxes. Finally, we also investigate regional heterogeneity given that most firms are located in the northern regions of Italy.

ABSTRACT. Inter-municipal cooperation to provide joint public good has been promoted in many European countries. We analyse coalition formation using a unique panel data of 1,056 municipalities within the French region of Brittany. We use a control function approach to develop a binary discrete choice model with spatial interaction. Imposed national tax limit is exploited as exclusion restriction to estimate the decision to cooperate in response to neigh-boroughs decision. The comparison with spatial econometrics models (SAR and Durbin) shows that the decision of cooperation, however dependent on the decisions and the characteristics of neighbours jurisdictions, is over estimated by this more traditional models. These results are in line with the recent applied spatial economics literature as in Lyytikainen (2012) but are derived for a discrete choice model setting.

ABSTRACT. This paper investigates the early steps of Italian industrialization at the local level. The analytical framework is the conditional convergence model augmented to account for spatial effects. The geographic unit of analysis is constituted by provinces (NUTS 3 units) and the time period considered goes from 1871 to 1911. Our results suggest that education, a cooperative culture, and spatial spillovers are able to explain, other things being equal, much of the geographical variability of value added growth in the manufacturing industry in 19th century Italy.

ABSTRACT. This paper investigates whether the classic Mundell-Flemming "trilemma" has morphed into a "dilemma" due to the rise of financial globalisation since the late 1990s. According to this new "dilemma" view, global financial cycles determine domestic financial conditions regardless of an economy's exchange rate regime, and independent monetary policy is possible if and only if capital mobility is restricted. We estimate the output response to a domestic monetary policy shock for a sample of advanced and emerging market economies over 1999-2009 using a mixed cross-section global VAR model. We examine whether heterogeneities in domestic monetary policy transmission can be explained by cross-country differences in global financial integration patterns. We show that the impact of financial globalisation on monetary policy effectiveness depends on the relative importance of two salient features of global financial integration. One is the size of an economy's external balance sheet, which determines the extent to which global financial cycles influence domestic financial conditions and, thereby, reduce monetary policy effectiveness. The other is the currency composition of the external balance sheet, which may give rise to exchange rate valuation effects and strengthen domestic monetary policy effectiveness. As the latter can only play out if the exchange rate is flexible, the choice of the exchange rate regime is critical to retain monetary policy autonomy under capital mobility. Hence, the "trilemma" remains valid.

ABSTRACT. This paper intends to test if investors diversify the risk due to domestic economic fluctuations, by investing in foreign firms and let them to specialize and exploit comparative advantages. The greater specialization is then expected to increase the volatility of the owned firm. We empirically assess to which extent the volatility of firms is due to activities of firms under foreign ownership, both controlling for idiosyncratic risk diversication, physical distance between shareholding/controlled firms and domestic/ foreign investors' prole. Following, Kalemli-Ozcan, Sørensen, and Volosovych (2010) we provide empirical evidence that risk sharing increases the volatility of controlled firms and therefore enhances specialization in production. To the best of our knowledge, this well-established and important theoretical proposition has not been tested before at so disaggregated level. To empirically test the diversification hypothesis, we build a novel dataset of micro (at firm level) and macro data (at the level of states and regions) composed by 400.000 large and very large firms in the EU over the years 19852012, for a total of approximately 2.9 million observations. By calculating a time-varying measure of firm volatility based on three indicators (turnover, sales and number of employees) we show how the lower is the degree of covariation of GDP growth among the economies of the shareholding/owned firms, the higher is the volatility of firms. Moreover, there is a direct and proportional link between the degree of volatility and the physical distance between the two firms. Finally, we do not find evidence that also domestic investors invest in firms located in other regions by looking for diversifying their risk and hedge against home regional business cycle effects. This suggests that only foreign shareholders are more prone to undertake risky investments and to force firms to specialize more when they invest abroad. These results are valid even considering several robustness checks in terms of different measures of volatility, types of firms (listed, export oriented, etc.), and location of shareholders.

ABSTRACT. A growing literature suggests that institutional quality is an important determinant of trade flows. Theoretical models of international trade with incomplete contracts predict that firms will not source intermediate inputs signing arm’s length contract if ex-post they cannot enforce the contract. Contract enforcement ultimately depend on the quality of institution. Prior empirical evidence support this idea using cross country data. We study how institutional quality at the local level influences the ability of firms to become international subcontractors using firm-level data. Using a sample of Italian firms and the trial length of civil disputes to proxy for contract enforcement, we find that firms located in courts with higher duration of processes have a lower probability to supply customized intermediate inputs to foreign firms.

ABSTRACT. This paper investigates the role of higher order beliefs in the formation of exchange rates. Our model combines a standard macroeconomic dynamics for the exchange rates with a microeconomic specification of agents’ heterogeneity and their interactions. The empirical analysis relies on a state space model estimated through Bayesian methods. We exploit data on macroeconomic fundamentals in a panel of subjective forecasts on the euro/dollar exchange rate. The equilibrium strategy on the optimization process of the predictors shows that higher order beliefs is the relevant factor in performing individual forecasting. Moreover public information, namely past exchange rates and fundamentals, plays a crucial role as a coordination device to generate expectations among agents on the basis of their forecasting abilities.

ABSTRACT. This paper analyses the relationship between the availability of low cost childcare services and maternal labour supply in Italy. By means of a simple job search model we show that lowering the price of childcare not only boosts maternal labour market participation, but may also raise the probability of getting a job by decreasing the reservation wage for which mothers are willing to accept a job offer. Exploiting discontinuities in the rules that determine children's eligibility to different types of childcare in Italy, we test the predictions of our model and evaluate the effects of a policy reform which expanded access to public kindergarten. Our results show that the possibility of anticipating the child's entry to kindergarten, a much cheaper alternative to daycare, increased both participation to the labour market and the employment probability of mothers. We show that the latter effect is largely driven by a decrease in the stated reservation wage.

ABSTRACT. The Italian General Government has the largest amount of trade debts and the longest payment delays in the Euro Area. This has serious negative effects on the financial situation of firms having commercial relations with it. Since 2012, the Government has made several steps to hasten firms' reimbursements, that culminated with making available in 2013, mostly to local governments, about 25 billion euros, to repay trade debts due at the end of 2012. By using a composite dataset, we evaluate the effect of these policies on a wide array of firms' financial and real business cycle indicators, distinguishing between firms that were repayed thanks to the measures, those that were not repayed although they had a legitimate right and those that did not have commercial relations with the General Government. Our results suggest that receiving money had a significant positive impact on firms' financial position, reducing their needs to assign credits to financial intermediaries and the probability of default on part (or all) of their debts. To some extent, also firms not receiving money benefited from the policies. As for the other non-financial indicators, we estimate a positive effect only on their investment in 2013 compared with initial plans, but not on the other indicators, possibly because of the short time elapsed between the repayments and the survey that collected the data.

ABSTRACT. This paper investigates the effect of government fragmentation on government spending using data on Basque municipalities' political and fiscal outcomes over four electoral terms. To identify causal errects, I use a natural experiment determined by the ban of a political party due to its tolerance to terrorism. I exploit the heterogeneity of the impact of the ban across municipalities to construct instruments for absolute majorities based on the mechanical changes in absolute majorities due to the ban. I find that absolute majorities reduce current expenditures, mainly by reducing spending on public goods and services. I don't nd political competition to have any effects on policy.

ABSTRACT. This paper studies the identication and estimation of treatment response with heteroge- neous spillovers in a network model. We generalize the standard linear-in-means model to allow for multiple groups with between and within-group interactions. We provide a set of identication conditions of peer effects and consider a 2SLS estimation approach. Large sample properties of the proposed estimators are derived. Simulation experiments show that the estimators perform well in nite samples. The model is used to study the effectiveness of policies where peer effects are seen as a mechanism through which the treatments could propagate through the network. When interactions among groups are at work, a shock on a treated group has effects on the non-treated. Our framework allows for quantifying how much of the indirect treatment effect is due to variations in the characteristics of treated peers (treatment contextual effects) and how much is because of variations in peer outcomes (peer effects).

ABSTRACT. We derive a closed-form expression for the posterior probability of Granger-noncausality in a Gaussian vector autoregression with a conjugate prior. We also express in closed form the posterior probability of Granger-causal-priority, a more general relation that accounts for indirect effects between variables and therefore is suitable in a multivariate context. We show how to use these results to choose variables for a vector autoregression, whether the goal is prediction or impulse response analysis.

ABSTRACT. We make available simple and accurate closed-form approximations to the marginal distribution and to the higher-order moments of Markov-switching Vector Auto-Regressive (MS VAR) processes. Our approximation is built upon the property of MS VAR processes of being gaussian conditionally on semi-innite sequences of the latent state. Assuming that normality holds given latent sequences of nite length and averaging over all possible sequences yields a mixture of normals that converges to the unknown marginal distribution as the sequence length increases. Numerical experiments conrm the viability of the approach. The same device applies to the calculation of the higher-order moments which were so far unavailable in the multivariate MS framework in spite of their relevance. Further applications to the closely related MS state space models are discussed.

ABSTRACT. This paper proposes a graphical model approach to the curse of dimensionality problem in high-dimensional multivariate autoregressive (MAR) models. We discuss several Bayesian learning algorithms that allows for local sparsity constraints to control the number of predictors of the model for parsimony. We also discuss joint inference of the temporal dependence in the observed time series and the maximum lag order of the process based on nested and non-nested graphical models. We investigate the performance and efficiency of the various algorithms on simulated and empirical experiments. The applied contribution compares our graphical approach with the Lasso-based model as a benchmark for modeling and forecasting selected macroeconomic time series with many predictors. The result shows a higher predictive performance and parsimony in favor of our model approach over the Lasso-based model for high-dimensional time series.

ABSTRACT. It may be desirable, for various reasons, to establish criteria for SVAR identification that do not depend on unknown parameters, but only on the set of restrictions that are imposed on the system a priori on theoretical grounds.

In the context of linear systems, this was accomplished in Johansen (1995). This paper extends and amends the approach proposed by Lucchetti (2006); we introduce a set of criteria which ensure identification independently of unknown parameters for a reasonably general class of models and discuss its possible generalisation.