ABSTRACT. Surprises in the price of oil futures computed on the day of OPEC announcements have been used as an exogenous measure of shifts in market beliefs about future oil supply to identify shocks to oil supply expectations. I show that these surprises capture not only revisions in market expectations about oil supply, but also revisions in expectations about oil demand. This conflation of supply and demand components invalidates the use of the surprises as an exogenous measure of shocks to oil supply expectations. I show that imposing an additional restriction on the sign of the co-movement between surprises in oil futures and changes in stock prices within the day of the OPEC announcement disentangles the two underlying shocks. Accordingly, I derive two robust instruments for the identification of shocks to oil supply and demand expectations that combine the surprises with this additional sign restriction, and I test them on a set of empirical specifications modelling the oil market and the global economy. A negative shock to oil supply expectations has deep and long-lasting stagflationary effects on global economic conditions that are stronger and more immediate than previously reported. These effects represent a challenge for monetary authorities that seek to stabilise both prices and output. I show that information effects are a prominent feature of the oil market and need to be accounted for when estimating the effects of shocks to oil supply expectations.

ABSTRACT. We document the existence of a sizable and statistically significant effect of oil shocks on US banking variables by estimating a SVAR with sign restrictions as in Baumeister and Hamilton (2019). In general, contractionary oil shocks decrease banks’ net worth, increase the credit spread, and lower the amount of credit to the economy. In particular, economic activity shocks have long-lasting effects, especially on credit, while oil supply and consumption demand shocks have temporary, short-lived effects. Most oil shocks meaningfully contribute to explain banking variables dynamics during commonly studied periods of economic significance, e.g., the Great Recession or the 2014-2016 period.

ABSTRACT. In this paper, we aim at quantifying virtually any possible distributive effects that energy price shocks can exert on flow-variables, i.e. except for wealth. We look at how energy price shocks change the distribution of income at the macro level (e.g., by looking at the functional income distribution) and at the personal level. Moreover, we try to estimate how the distribution of profits across sectors, the distribution of consumption expenditures across goods and along income distribution, and consumption inequality are impacted by energy price shocks. We focus on the U.S. and employ a Proxy-SVAR using the OPEC announcements oil price instrument designed by Känzig (2021). We make use of aggregate (NIPA tables) and micro (Survey of Consumer Finances, SCF) data. Our main contribution to the literature is to estimate as many as possible distributional effects along any relevant dimension of inequality and possible transmission channel originated by energy price shocks.

ABSTRACT. This paper studies the identification of Structural VectorAutoregressions (SVARs) exploiting a break in the variances of thestructural shocks. Point-identification for this class of models relieson an eigen-decomposition involving the covariance matrices ofreduced-form errors and requires that all the eigenvalues are distinct.This point-identification, however, fails in the presence ofmultiplicity of eigenvalues. This occurs in an empirically relevantscenario where, for instance, only a subset of structural shocks had thebreak in their variances, or where a group of variables shows a varianceshift of the same amount. Together with zero or sign restrictions on thestructural parameters and impulse responses, we derive the identifiedsets for impulse responses and show how to compute them. We performinference on the impulse response functions, building on the robustBayesian approach developed for set identified SVARs. To illustrate ourproposal, we present an empirical example based on the literature on theglobal crude oil market where the identification is expected to fail dueto multiplicity of eigenvalues.

ABSTRACT. This paper studies the ESG impact to the downside risk of companies in the US market by introducing a novel measure, the ESG risk contribution (△CoESGRisk). △CoESGRisk is a measurement based on the co-movement between the ESG risk factor and the downside risk. When there is a sudden increase in the ESG risk factor, the downside risk of high-ESG companies is reduced. However, under extreme conditions, the downside risk of high-ESG companies could also be increased, due to the increased volatility. The ESG impact is positively correlated with the ESG performance and size, and it varies among sectors.

ABSTRACT. Energy transition requires huge amounts of critical metals — called energy transition metals (ETMs) — to deploy clean energy technologies. The growing demand for ETMs and uncertainties regarding the path to net zero emissions could cause wide ETM price oscillations, with potential effects on the prices of other commodities. We explore whether upward and downward movements in ETM prices have a neutral effect on the level and volatility of energy and non-energy commodity prices. By characterizing the conditional dependence between ETM and commodity prices, we document that, except for natural gas, extreme ETM price changes have a non-neutral effect on commodity prices, although this effect vanishes for non-extreme price movements. The implications of this evidence for investors operating in commodity markets are evaluated in terms of commodity risk-adjusted returns, commodity tail risk, and liquidity needs for trading in commodity futures contracts.

ABSTRACT. The growing attention on sustainability in economics and finance has prompted the use of Environmental, Social, and Governance (ESG) metrics by practitioners, academics, and investors. The sustainable finance literature has mainly focused on investigating inconsistencies across ESG ratings. In this paper, we assume a different position focusing on how much and what is in common among the ESG metrics. We assume that ESG metrics are combined by a structural nonrandom part and a random part, that interferes with the structure. Furthermore, we are interested in studying if the reference variables are comparable across different data providers. We build an extensive dataset gathered from three distinct commercial databases covering more than 5,000 listed companies at the worldwide level. As a preliminary analysis, we focus on the Environmental component of the ESG score, focusing on the subset of companies in our database operating within the Electricity sector, strongly impacted by environmental risks. Specifically, for each of the three databases, we apply the PCA on the indicators underlying the ‘Environmental’ score to investigate the main components that determine it. We identify five items consistently emerging within the main components across the databases. While confirming a significant disagreement among ESG raters, we highlight the emergence of some common factors that, alone, explain significant aspects of the phenomenon. We will extend this investigation by replicating this empirical approach across scores and industries. We will also provide an analysis in time-series in order to understand if the estimated loadings change over time.

ABSTRACT. An extended specification for estimating the risk premia necessary for the forward pricing of wholesale electricity is developed in order to respond to the increasing need for more precise risk management of hedging positions in practice. Using Taylor expansions, we provide new specifications for the electricity forward premium including its dependency on all four moments of the expected wholesale price density as well as the higher moments of the demand density including skewness and kurtosis. Overall we argue that previous models have been underspecified and that the extended formulation proposed in this analysis is robust and worthwhile.

ABSTRACT. We propose two specifications of a real-time mixed-frequency semi-structural time series model for evaluating the output potential, output gap, Phillips curve, and Okun's law for the US. The baseline model uses minimal theory-based multivariate identification restrictions to inform trend-cycle decomposition, while the alternative model adds the CBO's output gap measure as an observed variable. The latter model results in a smoother output potential and lower cyclical correlation between inflation and real variables but performs worse in forecasting beyond the short term. This methodology allows for the assessment and real-time monitoring of official trend and gap estimates.

ABSTRACT. The relationship between energy consumption and economic growth is debated in academic literature. As cleaner energy sources gain effectiveness in reducing global pollution, countries shift towards sustainable production. Energy regulations like the Kyoto Protocol and Paris Agreement aim to address climate change and promote renewable energy. This paper uses cointegration and Granger causality analysis on panels of high-income European countries, low-income European countries, America, and Asia. Results support the Environmental Kuznets Curve (EKC) for low-income EU countries and high-income European countries with nuclear energy. Heterogeneity exists in the relationship between air pollution and economic growth. Granger-causality reveals varying relationships between GDP and energy consumption: high-income EU and US countries exhibit the conservation hypothesis (GDP → energy), Asian countries show the growth hypothesis (Energy → GDP), and low-income EU states demonstrate the feedback hypothesis (GDP ↔ energy). We introduce the "decrease" hypothesis for low-income European countries, suggesting that unidirectional causality from renewable energy consumption to GDP leads to decreased economic growth. These findings have important implications for policymakers and governments.

ABSTRACT. On the 24th of February, Russia launched its military invasion of Ukrainian territories. Potentially, this event could impact the global economy harder than the COVID-19 outbreak, given the two countries’ role as exporters. This paper aims to analyze the effect of the invasion of Ukraine on the international stock markets, taking into account sustainability performance at the company and industry levels. We obtained two preliminary results. First, European countries were the most affected by the event of invasion. Secondly, on the day of the invasion and in the following days, Green companies showed larger Abnormal Returns (ARs) than those of the Brown companies.

ABSTRACT. We quantify the economic impact of Russia’s invasion of Ukraine on Germany, the United Kingdom, France, Italy and Switzerland using data on historical geopolitical events. Applying a structural VAR approach based on sign and narrative sign restrictions, we find that the war has exerted a notable drag on real activity and has pushed inflation up considerably. For example, a counterfactual exercise suggests that in Germany, GDP would have been 0.7 percent higher and the CPI 0.4 percent lower in 2022Q4 if Russia had neither attacked nor threatened Ukraine. The negative consequences of the war are likely to be far greater in the medium-to-long term, especially with regard to the real economy.

ABSTRACT. This paper investigates the recent increases in natural gas prices and its propagation effects via inflation expectations. Using a structural vector autoregression, we identify a euro area natural gas price shock with a combination of sign- and zero-restrictions. We rely on market-based measures of inflation expectations. We find that natural gas price shocks have strong effects on both inflation and inflation expectations. To understand the relative importance of the pass-through from inflation expectations to inflation after a natural gas price shock, we conduct a counterfactual analysis in which we turn off the expectation channel. Our findings indicate the presence of strong second-round effects via expectations. Furthermore, these effects are stronger for short-term expectations than for long-term expectations. Our analysis provides a guidance for policymakers to better understand the potential trade-offs of different policy responses to natural gas price shocks.

ABSTRACT. Over the last twenty years, temperature variability has been increasing across Europe, affecting the economy through the demand for energy for heating and cooling needs. This paper provides empirical evidence about an energy transmission channel of temperature anomalies to inflation by estimating a Structural Vector Autoregressive model (SVAR) for six Euro-Area countries. Starting from grid-level meteorological data, we find that temperature anomalies negatively affect energy demand, suggesting that a ”turn-off-heating” effect outweighs the ”turn-on-cooling” in Europe. Furthermore, the first fundamental law of thermodynamics, which establishes a relationship between temperature and vapour pressure, allows us to set a novel sign-restriction identification that distinguishes warm spells (positive anomalies) from cold ones (negative anomalies). We find that the former affects energy prices in several countries, while the latter features mixed effects. Overall, the temperature shock dampened price growth in the EA: in 2015-2021, the annual energy and the headline inflation were lower by 0.15% and 0.05%, respectively, due to temperature anomalies. Those results shed new light on the effect of global warming on the EA economy and may have implications for the monetary authority.

ABSTRACT. Shocks to European gas prices likely have a different effect on inflation depending on what drives the shock – supply issues, a higher need for energy as activity expands, or more demand for gas inventories as a precaution. In this paper, we develop a Bayesian VAR model to identify different types of gas price shocks in the European gas market and we document how these feed through to prices in the euro area. Our estimates show that the pass-through of gas shocks is larger than its weight in the expenditure basket and can be persistent over time, depending on the type of shock. We show that gas supply shocks pass through all components of euro area inflation, to producer prices, wages and core inflation, and that the response is significantly magnified in periods of low unemployment.

ABSTRACT. We develop a Bayesian Structural VAR model to study the relationship betweendifferent energy shocks and inflation dynamics in Europe. Specifically, we model theendogenous transmission from shocks identified by the global market of crude oiland the European natural gas market to two target macroeconomic variables, i.e.inflation expectations and realized headline inflation rate. Our results demonstratethat, since the post-pandemic recovery, inflation in the Euro area is mostly drivenby energy price shocks and aggregate supply factors. In particular, the high peaksof the Eurozone inflation are mostly associated with natural gas supply shocks

ABSTRACT. We introduce a new measure of intermittency, given by realized volatility of wind speed based on high-frequency (hourly) observations, and show that it comes through as a strongly significant explanatory variable in an analysis of the dynamic relation between CO$_2$ emissions, net electricity import, and wind energy in Denmark. Intermittency reduces wind generation and increases emissions. As the system variables are strongly persistent and move together in the long run, we pursue a fractional cointegration approach, extended to accommodate covariates, including intermittency, climate, and demand variables. Temperature reduces emissions, precipitation increases wind generation, and wind speed increases wind generation, while reducing electricity import. Marginal emissions avoided (MEA) are estimated at 0.53 tonnes per MWh of wind generation and significant, based on the bootstrapped confidence interval, hence confirming the emissions abatement potential of wind energy adoption in the presence of intermittency.

ABSTRACT. It is still unclear to what extent transition risks are being internalised by financial investors. In this paper, we provide a novel investigation of the impact of media-based measures of transition risks on credit risk of energy companies, as measured by their CDS indices, in both Europe and North America. Using linear and non-linear local projections, we find that, in both jurisdictions, a transition risk shock affects CDS indices only when combined with tangible physical climate-related impacts. We also find evidence of non-linear cross-border effects, with North American energy companies particularly affected by European dynamics. We suggest that the public reaction in the wake of severe natural disasters, which might push policy-makers to adopt more decisive climate action, contribute to making the transition-related debate salient in the eyes of credit market actors.

ABSTRACT. Understanding the economic impact of weather events, such as those that science is increasingly linking to climate change, is crucial for policy-making and to design effective damage mitigation strategies. In this paper, we study the impacts of thunderstorms, which are less extreme than floods or hurricanes, but can still be highly damaging and affect much broader regions. We analyze their potential impacts on income and wages using a comprehensive panel dataset spanning three decades, and capturing more than 200,000 storm events of varying strength in the United States. Our findings reveal a significant and highly robust negative association between storm activity and income and wages. Notably, while income tends to recover in the long run, wages exhibit a more stubborn decline, suggesting persistent impacts on income inequality. Our analyses also highlight a lack of effective hazard-driven adaptation and the existence of significant adaptation gaps, with economically disadvantaged areas displaying stronger negative impacts. Moreover, we find evidence for an important role of federal assistance and support, which effectively counteract storm-induced losses. Given that climate change is likely to increase storms' intensity and magnitude, our results show the need for comprehensive policies to address the complex dynamics of storm-induced impacts, close existing adaptation gaps and ensure an equitable outcome in presence.

ABSTRACT. We discuss a new way to construct Regime Switching Autoregressions making use of a non-Markovian unobserved process. We show that, in a special case, the likelihood implied by this new specification is identical to the classical Markov Switching Autoregression one. The general case leads to more flexible specifications with tractable likelihood functions. We discuss the statistical properties of the model and establish conditions for the consistency and asymptotic normality of the Maximum Likelihood Estimator. An application shows that the new specification leads to better estimates and predictions.

ABSTRACT.  When in proxy-SVARs the covariance matrix of VAR innovations is subject to exogenous, permanent (nonrecurring) breaks that generate target impulse response functions (IRFs) that change across volatility regimes, even strong, exogenous external instruments can result in inconsistent estimates if the breaks are not properly accounted for. In such cases, it is essential to explicitly incorporate the shifts in unconditional volatility in order to point-identify the target structural shocks and possibly restore consistency. We show that if a necessary and sufficient rank condition that exploits the moments implied by the changes in volatility holds, the target IRFs can be point-identified and estimated consistently. Importantly, standard asymptotic inference applies despite (i) the covariance between the proxies and the instrumented structural shocks being local-to-zero as in Staiger and Stock (1997), (ii) instruments exogeneity possibly fails. It turns out that external instruments never compromise the inference on the dynamic causal effects exerted by the target shocks. In the worst-case scenario, they merely serve as labels for the target structural shocks. We present a novel identification strategy that appropriately combines external instruments with "informative" changes in volatility regimes, thereby avoiding the need to assume proxy relevance and exogeneity in estimation. We illustrate the usefulness of the suggested method by reexamining some proxy-SVARs previously estimated in the existing literature, focusing in particular on the identification and estimation of US fiscal multipliers by fiscal external instruments.

ABSTRACT. State estimation in unobserved component models with parameter uncertainty is traditionally performed through approximate filters based on collapsing schemes, where Gaussian distributions with given moments are employed to replace otherwise intractable conditional densities. This paper considers a signal-plus-noise model where parameter uncertainty is induced by a latent variable that may assume a fixed number of states. We show that for these models, the approximate filter commonly adopted coincides with the one implied by a dynamic adaptive mixture model, where the parameters of a mixture of distributions evolve over time following a recursion that is based on the score of the one-step-ahead predictive distribution. We focus on a robust specification, where the mixture components are Student-t distributions. The stochastic properties of the filter and asymptotic properties of the maximum likelihood estimator of the static parameters are derived. An application to the US industrial production index, where the novel specification is compared with the alternative class of mixture autoregressive models, is provided. A second application to daily volume of GAS shows the case when Student-t mixtures outperform Gaussian mixtures.

ABSTRACT. In production theory, conditional frontiers and conditional efficiency measures area flexible and appealing approach to consider the role of environmental variables onthe production process. Direct approaches estimate non-parametrically conditionaldistribution functions requiring smoothing techniques and the use of selected bandwidths.The statistical literature produces way to derive bandwidths of optimal order,by using e.g. least-squares-cross-validation techniques. However, it has been shownthat the resulting order may not be optimal when estimating the boundary of thedistribution function. As a consequence the direct approaches may suffer from somestatistical instability. In this paper we suggest a full nonparametric approach whichavoids the problem of estimating these bandwidths, by eliminating in a first step theinfluence of the environmental factors on the inputs and the outputs. By doing thiswe produce “pure” inputs and outputs which allow to estimate a “pure” measure ofefficiency, more reliable for ranking the firms, since the influence of the external factorshave been eliminated. This can be viewed as an extension of the use of location-scalemodels (implying some semi-parametric structure) to full nonparametric models, basedon nonseparable, nonparametric models. We are also able to recover the frontier andefficiencies in original units. We describe the method, its statistical properties and weshow in some Monte-Carlo simulations, how our new method dominates the traditionaldirect approach.

ABSTRACT. This paper studies the problem of assessing the predictive accuracy of competing forecasts when the target variable is observed with error. The theoretical and empirical literature on forecasting accuracy has invariably used unbiased proxies. After introducing the notion of proxy reliability, which does not necessitate unbiasedness, we show how to optimally use biased proxies to maximize the probability of inferring the true (unknown) ranking. Our procedure still preserves the robustness of the loss function, in the sense of Patton (2011b), and allows testing for equal predictive accuracy, as in Diebold and Mariano (1995). We illustrate the potential of the procedure by evaluating forecasts for GDP growth and financial market volatility.

ABSTRACT. Accurate short-term now- and forecasting of global temperatures is an important issue and helpful for policy design and decision making in the public and private sector. We compose a raw mixed-frequency data set from weather stations around the globe (1920-2020). First, we document smooth variation in average monthly and annual temperature series by applying a dynamic stochastic coefficient model. Second, we use adaptive cross-validated forecasting methods which are robust to smooth changes of unknown form in the short-run. Therein, recent and past observations are weighted in a mean squared error-optimal way. Overall, it turns out exponential smoothing methods (with bootstrap aggregation) often performs best. Third, by exploiting monthly data, we propose a simple procedure to update annual nowcasts during a running calendar year and demonstrate its usefulness. Further, we show that these findings are robust with respect to climate zones. Finally, we investigate now- and forecasting of climate volatility via a range-based measure and a quantile-based climate risk measure.

ABSTRACT. I use Bayesian VARs to forecast global temperatures anomalies until the end of the XXI century by exploiting their cointegration with the Joint Radiative Forcing (JRF) of the drivers of climate change. Under a ‘no change’ scenario the land temperature anomaly is projected to reach 6.9 Celsius degrees in 2100. Forecasts conditional on alternative paths for the JRF show that, given the extent of uncertainty, bringing climate change under control will require to bring the JRF back to the level reached in the early years of the XXI century. In recent years the JRF has exhibited a marked acceleration, which has not been reflected yet in temperature anomalies, and points towards their corresponding acceleration going forward. From a methodological point of view, my evidence suggests that previous cointegration-based studies of climate change suffer from model mis-specification.

ABSTRACT. Tipping points are permanent environmental changes that cannot be undone even after implementing GHG abatement policies. This paper proposes an index of environmental change that might be useful to measure the likelihood of the occurrence of such irreversible events given current trends or preemptive abatement scenarios. An application to European data illustrates its construction and implementation for forecasting analysis.

ABSTRACT. This book presents the fundamental theoretical notions of supervised machine learning along with a wide range of applications using Python, R, and Stata. It provides a balance between theory and applications and fosters an understanding and awareness of the availability of machine learning methods over different software platforms. After introducing the machine learning basics, the focus turns to a broad spectrum of topics: model selection and regularization, discriminant analysis, nearest neighbors, support vector machines, tree modeling, artificial neural networks, deep learning, and sentiment analysis. Each chapter is self-contained and comprises an initial theoretical part, where the basics of the methodologies are explained, followed by an applicative part, where the methods are applied to real-world datasets. Numerous examples are included and, for ease of reproducibility, the Python, R, and Stata codes used in the text, along with the related datasets, are available online. The intended audience is PhD students, researchers and practitioners from various disciplines, including economics and other social sciences, medicine and epidemiology, who have a good understanding of basic statistics and a working knowledge of statistical software, and who want to apply machine learning methods in their work.

ABSTRACT. This paper is joint with James Hamilton.

Modeling how individual units interact to determine aggregate outcomes can be a rich source of identifying information. We use this insight to develop a generalization of granular instrumental variables estimation and show how parameters of a dynamic structural model can be estimated using full-information maximum likelihood. We apply the method to a study of the world oil market. We conclude that the supply responses of Saudi Arabia and adjustments of inventories have historically played a key role in stabilizing the price of oil.

ABSTRACT. This paper shows that the impact of the global money supply is disproportionallyhigh for energy than for non-energy commodities prices. An increase in the globalmoney supply for energy commodity prices results in demand-pull inflation. However,for non-energy commodity prices, it leads to demand-pull and cost-push inflation, asenergy is a key input for non-energy commodities. To quantify this effect, we useda Markov switching model with time-varying transition probabilities. The modelconsiders periods of slow, moderate, and fast global money supply growth. We findthat the response is almost double for energy than for non-energy commodity prices.

ABSTRACT. The EU carbon market serves as an innovative financial instrument with the primary objective of contributing to mitigate the impacts of climate change. This market demonstrates significant interconnectedness with fossil energy, precious metal, and financial markets, although limited research has focused on the causality, dependency, intensity and direction of time-varying spillover effects. This study aims to investigate the causality direction, degree of dependency structure, and volatility transmission from Brent Oil, UK Natural Gas, Rotterdam Coal, Gold, and EuroStoxx600 future prices to EU Allowances during various crisis periods. The following aspects of these relationships are examined: causality direction across markets, dependency structure of their joint distribution, and the information transmitted among markets. To achieve these objectives, this paper proposes a novel methodological approach that combines the most recent econometrics methods, such as Directed Acyclic Graph analysis, C-Vine Copula models, and Time-Varying parameter Vector AutoRegressive models with Stochastic Volatility with the use of a comprehensive sample of daily data from 26 April 2005 to 31 December 2022. The major findings of this study demonstrate that causality predominantly runs from energy, metal, and financial markets to the EU carbon market. The dependency structure, although varying across different sub-periods, shows a strong relationship observed between fossil fuels, particularly coal, and CO2 market. Additionally, the EuroStoxx600 futures price index exhibits the highest dependence on EUA prices. Furthermore, the study establishes that the EU carbon market is a net receiver of shocks from all other markets, with the energy and financial markets significantly influencing volatility in EUA prices. The time-varying spillover effect is most pronounced with a one-day lag, and the duration of the spillover effects between carbon and EuroStoxx600 ranges from 3 to 15 days, gradually diminishing over time. These results have the potential to increase the understanding of the EU carbon market and offer practical guidance for policymakers, investors, and companies involved in this domain

ABSTRACT. We examine the impact of changes in selected critical raw materials consumptionand production resulting from the energy transition, in the US market. To achievethis, we estimate a Structural Vector Autoregressive model (SVAR) which disen-tangles demand- and supply- driven shocks. Additionally, we employ a structuralforecast analysis, up to 2030, to study price changes under dierent scenarios.By conditioning price forecasts to various demand and supply scenarios, we inves-tigate the implications of the US Ination Reduction Act (IRA) and the associatedpolicies aimed at boosting US domestic production of these critical minerals.Our results highlight how dierent minerals exhibit distinct dynamics, empha-sizing the need to treat them as separate entities rather than a homogeneous group.Moreover, we provide price forecasts which are sensitive to the chosen scenarios,hence oering valuable insights into critical minerals' price dynamics.

ABSTRACT. In this paper, we revisit early contributions by Diebold and Li (2006) and Diebold et al. (2006) in the light of more recent development in the literature on Dynamic Factor Models and provide evidence on how differences in the methods for extracting yield curve factors are reflected in their forecasting performance. The fact that interest rates normally behave as I(1) variables in most cases has been handled in DFMs either by ignoring the possible consequences of nonstationarity or by differencing the data. However, as shown by Casoli and Lucchetti (2021), unit roots can be accommodated if cointegration relationship between the observables are present.

ABSTRACT. We propose a per-cluster instrumental variables approach (PCIV) for estimating linear correlated random coefficient models in the presence of contemporaneous endogeneity and two-way fixed effects. This approach estimates heterogeneous effects and aggregates them to population averages. We demonstrate consistency, showing robustness over standard estimators, and provide analytic standard errors for robust inference. In Monte Carlo simulation, PCIV performs relatively well in finite samples in either dimension. We apply PCIV in estimating the price elasticity of gasoline demand using state fuel taxes as instrumental variables. We find significant elasticity heterogeneity and more elastic gasoline demand on average than with standard estimators.

ABSTRACT. The margin between the price of crude oil and one or more of its refined products, known as the crack spread, is a key risk measure in the energy trading industry. Effective forecasts of crack spread returns require modelling the co-movements of crude oil and refined product prices. To this end, in this paper we re-examine the relationship between parameter instability in asset returns correlations and macroeconomic fundamentals using a new correlation component model dubbed the Regime Switching DCC-MIDAS (RSDCC-MIDAS), which distinguishes regime switches in the short and long-run correlations. Breaks in the secular component are associated with low-frequency macroeconomic fundamentals via a Smooth Transition MIDAS regression, while short-run correlations are characterized by abrupt breaks linked to market constraints. Following a discussion of estimation, inference and simulation-based evaluations, the model is applied to the prediction of energy futures returns. Results from an extensive forecasting exercise reveal the benefits of our specification in terms of forecasting performance at medium and long horizons and in times of intense market instability, such as the recent pandemic crisis; optimal portfolio allocation; and risk management strategies in crude oil and distillate markets.

ABSTRACT. The recent drops in oil prices associated with the Covid pandemic and spikes withthe Ukraine war have highlighted the importance of risk assessment and tail events inoil markets. The common model for studying tail events is quantile regression, withBayesian VARs with stochastic volatility (SV) a serious competitor. Besides thesetwo models, in this paper we suggest the use of a (time-varying) mixture of a linearBVAR and a Bayesian Additive Regression Trees (BART) VAR, labelled mixBART,to model and forecast a set of oil market fundamentals and oil prices, with a specificinterest in predicting tail events. To further increase the model flexibility, the errorsfollow a factor stochastic volatility (FSV) model. Empirically, we find relevant nonlinearitieswhen modelling oil prices, in particular in the presence of extreme events,which can be hardly captured by linear models with stable parameters. Moreover,there are forecasting gains from using mixBART, with respect to quantile regressionand BVAR-SV, larger for tail than for point forecasts, and larger in the right thanin the left tail. Finally, mixBART performs well for monitoring and predictingoil prices also during the Covid pandemic and Ukraine war periods, permittingrather fast changes not only in the location but also in the shape of the predictivedistributions.

ABSTRACT. Since Russia's invasion of Ukraine, many countries have pledged to end or restrict their oil and gas imports to curtail Moscow's revenues and hinder its war effort. Thus, the European ministers agreed to trigger a cap on the gas price. To detect the importance of the price cap for gas, we provide a mixture representation for the gas price to detect the presence of outliers made by a truncated normal distribution and a uniform one. We focus our analysis on Germany and Italy, which are major Russian gas importers by exploiting the response of the different commodities to a gas shock through a Bayesian vector autoregressive (VAR) model. As a result, including a lower gas price cap smooths the impact of a gas shock on electricity prices, while not considering a price cap will increase exponentially this impact. Regarding the other commodities, gas shocks matter in the short and long run when a price cap is not considered

ABSTRACT. In this study, we use data on European stocks to construct a green-minus-brown portfolio hedging climate risk and to evaluate its performance in terms of cumulative expected and unexpected returns. More specifically, we estimate a Structural Panel VAR fitted to one month return and realized volatility computed for 40 constituents of a green portfolio (i.e., the low carbon emission portfolio monitored by Refinitiv) and for 41 constituents of a brown portfolio (underlying the Oil&Gas and Utilities industry sectors of the STOXX Europe 600). The common shocks underlying the cross-sectional averages, interpreted as portfolio shocks, are retrieved in a first stage of the analysis and they are used to control for cross-sectional dependence. We compute the historical decomposition (for cumulative returns) in a second stage of the analysis and we find, in line with Pastor et al. (2022), an out-performance of the expected component of the brown portfolio relative to the one for the green portfolio, and an out-performance of the green portfolio when we turn our focus on the unexpected component. We also extend the analysis of Pastor et al. (2022), assessing, for the top 5 constituents of the green portfolio (e.g., those which are found to have the worst performance in terms of expected return), the role played by idiosyncratic shocks in shaping their out-performance in terms of unexpected component. Finally, after exploiting the non-gaussian time series properties of the financial time series considered for the purpose of statistical identification, we are able to interpret ex post the idiosyncratic shocks in terms of financial leverage and risk aversion.

ABSTRACT. This paper uses a time-varying Factor Augmented VAR model to construct a novel measure of climate uncertainty that is common to a set of fifteen advanced economies. We use this measure to investigate its evolving transmission on a range of macroeconomic and financial variables. With an aim to highlight its policy relevance, we simultaneously construct and compare its impact to a measure of common macroeconomic and financial uncertainty shock. Our findings have important policy implications as we find that the common climate uncertainty shock has contractionary properties, similar in spirit to the common macroeconomic and financial uncertainty shock. We find that the impact of both uncertainty shocks has declined systematically over time, albeit by different magnitudes. Our results suggest that while the negative impact of the common macroeconomic and financial uncertainty shock dominated the negative responses at the beginning of the sample, the common climate uncertainty shock dominated the negative responses of real activity and financial variables towards the end of the sample.

ABSTRACT. Using a monthly global vector autoregressive (GVAR) model at the State level for the United States, we estimate the effects of severe weather shocks on local economic activity and cross-border spillover effects operating through economic linkages between States. We measure shocks using a detailed county-level database on major natural disasters. Impulse response analysis shows significant country-wide macroeconomic effects of weather shocks hitting individual regions. We also show that (i) taking into account economic interconnections between States allows to capture much stronger spillover effects than those associated with mere spatial adjacency, (ii) geographical disaggregation and parameter heterogeneity across States are critical for assessing country-wide effects of weather shocks, and (iii) network effects amplify the local impact of these shocks.

ABSTRACT. We investigate the environmental Impacts – measured as CO2 emissions – of Population,Affluence and Technology (IPAT) for the major emitter countries at worldlevel. In this respect, we develop a global VAR model that: i) accounts for endogenousrelationships among all the IPAT variables; ii) encompasses cross-countriesspillover effects; iii) embodies structural scenarios and conditional forecasting analysisto assess the impacts of the “Net Zero Emissions by 2050” policy. Our results willbe crucial to quantify the role of international spillover and to evaluate alternative(“parallel”) PAThs that are compatible with a reduction in global CO2 emissions.

ABSTRACT. We arrive at this conclusion by using a new family of models—the Long Memory Dynamic Judgmental Protocol (LMDJP)—where robust filtering and fractionally integrated autoregressions are combined in an environment characterized by several players—namely, Forecast Producer, Forecast User and Reality. Our simulated and empirical evidence reveals that (i) knowledge of the LM parameter matters for the p-values of tests for spurious long memory; (ii) secondly that the role of long memory in belief formation is ambiguous.

ABSTRACT. The availability of data on economic uncertainty sparked a lot of interest in models that can timely quantify episodes of international spillovers of uncertainty. This challenging task involves trading off estimation accuracy for more timely quantification. This paper develops a local vector autoregressive model (VAR) that allows for adaptive estimation of the time-varying multivariate dependency. Under local, we mean that for each point in time, we simultaneously estimate the longest interval on which the model is constant with the model parameters.

The simulation study shows that the model can handle one or multiple sudden breaks as well as a smooth break in the data. The empirical application is done using monthly Economic Policy Uncertainty data. The local model highlights that the empirical data primarily consists of long homogeneous episodes, interrupted by a small number of heterogeneous ones, that correspond to crises. Based on this observation, we create a crisis index, which reflects the homogeneity of the sample over time. Furthermore, the local model shows superiority against the rolling window estimation.

ABSTRACT. This paper proposes a novel time-series model with a non-stationary stochastic trend, locally explosive mixed causal non-causal dynamics and fat-tailed innovations. The model allows for a description of financial time-series that is consistent with financial theory, for a decomposition of the time-series in trend and bubble components, and for meaningful real-time forecasts during bubble episodes. We provide sufficient conditions for strong consistency and asymptotic normality of the maximum likelihood estimator. The model-based filter for extracting the trend and bubbles is shown to be invertible and the extracted components converge to the true trend and bubble paths. A Monte Carlo simulation study confirms the good finite sample properties. Finally, we consider an empirical study of Nickel monthly price series and global mean sea level data. We document the forecasting accuracy against competitive alternative methods and conclude that our model-based forecasts outperform all these alternatives.

ABSTRACT. We present a novel procedure, namely the LASSO-based Outlier Detection and Estimation (LODE) algorithm, to detect additive outliers and structural breaks (e.g., level shift) in structural time series within a State Space representation framework. Consider a structural time series given by the sum of three unobservable components: trend, seasonality, and a Gaussian additive error term. Also, assume that trend and seasonality are stochastic processes with zero-mean Gaussian disturbances having unknown variance to be estimated. The algorithm relies on a simple additive decomposition of the level's variance (for level shifts) and the disturbance's variance (for additive outliers) into a constant part and a time-varying part. Candidate change points identification and parameters estimation issues are jointly addressed within a penalized likelihood framework in which the common variance is an unknown parameter to be estimated but not penalized, while the additional parameters are penalized through a LASSO penalty capable of shrinking unneeded coefficients at zero. Inference for the parameters is performed by maximum likelihood, while optimal shrinkage parameters are obtained using Bayesian Optimization algorithms for likelihood criteria. Eventually, LODE is applied to study the evolution of fuel consumption across Italian provinces following the mobility restrictions imposed during the COVID-19 pandemic.

ABSTRACT. We propose a family of test statistics to detect the presence of changepoints in the mean of dependent, possibly multivariate, functional-valued observations. Our statistics are based on a generalisation of the empirical energy distance; in particular, we propose a weighted version of it, which is designed in order to enhance the ability to detect breaks occurring at sample endpoints. The limiting distribution of the maximally selected version of our statistics requires only the computation of the eigenvalues of the covariance function, thus being readily implementable in the most commonly employed packages, e.g. R. We show that, under the alternative, our statistics are able to detect changepoints occurring even very close to the beginning/end of the sample. In the presence of multiple changepoints, we propose a binary segmentation algorithm to estimate the number of breaks and the locations thereof. Furthermore, we show that our statistics can be readily generalised to testing for distributional changes, by applying them to the empirical characteristic function. Simulations show that our procedures work very well in finite samples. We complement our theory with an application to temperature data.

ABSTRACT. Threshold autoregressive moving-average (TARMA) models extend the popularTAR model and are among the few parametric time series specifications to includea moving average in a non-linear setting. The state dependent reactions to shocksis particularly appealing in Economics and Finance. However, no theory is currentlyavailable when the data present heavy tails or anomalous observations. Here we providethe first theoretical framework for robust M-estimation for TARMA models and studyits practical relevance. Under mild conditions, we show that the robust estimator forthe threshold parameter is super-consistent, while the estimators for autoregressive andmoving-average parameters are strongly consistent and asymptotically normal. TheMonte Carlo study shows that the M-estimator is superior, in terms of both bias andvariance, to the least squares estimator, which can be heavily affected by outliers. Thefindings suggest that robust M-estimation should be generally preferred to the leastsquares method. We apply our methodology to a set of commodity price time series; therobust TARMA fit presents smaller standard errors and superior forecasting accuracy.The results support the hypothesis of a two-regime non-linearity characterised by slowexpansions and fast contractions.

ABSTRACT. Estimated Impulse Response Functions (IRFs) in nonlinear VARs are often used to investigate the propagation effects of uncertainty, monetary policy, and oil shocks.  However, nonlinearities in VAR models introduce complexities in defining and estimating IRFs, making it crucial to explore different approaches. We derive possible definitions of IRFs in the context of nonlinear structural VARs and show how different definitions lead to different possible estimators with potentially very different implications for policy analysis. The widely used time Varying VAR is used as a running example. We provide practical guidance on consistently estimating these IRFs, aiming to bridge the gap between theoretical developments and applied research.

ABSTRACT. This paper develops two fully modified OLS (FM-OLS) type estimators for systems of seemingly unrelated cointegrating polynomial regressions with common regressors, i.e., systems of regressions that include deterministic variables, integrated processes, integer powers of integrated processes as well as common - across (potentially subsets of) equations - integrated processes and integer powers of common integrated processes as explanatory variables. The stationary errors are allowed to be serially correlated and the regressors to be endogenous. Furthermore, the errors and regressors are allowed to be dynamically cross-sectionally correlated. The developed estimators have zero mean Gaussian mixture limiting distributions that allow for asymptotic normal or chi-squared inference. We use the Wald-type hypothesis tests as basis to formulate tests for general forms of group-wise poolability. In case group-wise poolability is not rejected, we provide the corresponding group-wise pooled variants of the developed FM-OLS-type estimators. Our simulations indicate that appropriate pooling leads, as expected, to improved performance of both the estimators and hypothesis tests based upon them.

ABSTRACT. Despite climate change and the energy transition towards renewable energy, petroleum oil remains a key component of the energy mix of most countries in the world. While the understanding of oil prices, its impact on the real economy and forecasting oil prices are well understood model-based forecasts of oil consumption and production are less researched. We aim to close this gap by forecasting yearly oil consumption from the Energy Institute's Statistical Review of World Energy applying several methods. First, we apply SV-BVAR and the multi-dimensional approach from Baumeister, Korobilis and Lee (2022) and apply it to the Statistical Review data. The forecasts are then compared with the following alternatives: the Unobserved Components Stochastic Volatility (UC-SC; Stock and Watson 2007, JMCB), three-pass regression filter (3PRF; Kelley and Pruitt 2015, JoE) and Dominant Drivers by Machine Learning (DDML; Ditzen and Ravazzolo 2022).

ABSTRACT. This study examines short-to-mid-term point forecasting of daily electricity prices, with particularfocus on the role of renewable sources. Using data from the market zone corresponding to thenorthern region of Italy, we develop time series models with and without exogenous variables. Theoutput highlights the importance of load, weighted import price and natural gas price as influentialfactors in forecasting electricity price. Also, it supports the modest predictive influence of solarand wind generations, even considering their small contribution to the energy mix in northern Italy.However, hydropower generation indicates no predictive power for daily electricity priceforecasting. These outputs remain consistent whether using actual or forecasted exogenousvariables. These findings provide insight into potential outcomes if the contribution of renewablesources to the energy mix continues to grow.

ABSTRACT. Are energy firms more prone to corporate failure than non-energy firms? This paper investigates the risk of corporate failure in energy firms compared to non-energy firms in the U.S. through a survival analysis using the Cox proportional hazard model. Our findings reveal that energy firms face significantly higher failure risks than non-energy firms, independent of firm-specific variables, macroeconomic and market structures, and environmental factors. Even after accounting for financial statements and environmental variables, the risk of failure remains greater for U.S. energy firms. Additionally, macroeconomic variables have a more pronounced impact on the failure risk of energy firms, while environmental indicators play a more significant role in non-energy firms. The study also highlights the predictive power of the ESG score indicator for corporate failure. Although both energy and non-energy firms are sensitive to accounting indicators and ESG scores, there are notable differences in the magnitude and direction of these sensitivities between the two firm types.

ABSTRACT. In this paper, we forecast euro area inflation and its main components, remarkably energy, using an econometric model which exploits a massive number of time series on survey expectations for the European Commission's Business and Consumer Survey. To make estimation of such a huge model tractable, we use recent advances in computational statistics to carry out posterior simulation and inference. Our findings suggest that the inclusion of a wide range of firms and consumers' opinions about future economic developments offers useful information to forecast prices and assess tail risks to inflation. These predictive improvements do not only arise from surveys related to expected inflation but also from other questions related to the general economic environment. Finally, we find that firms' expectations about the future seem to have more predictive content than consumer expectations.

ABSTRACT. We introduce a new HD DCC-HEAVY class of hierarchical-type factor models for conditional covariance matrices of high-dimensional returns, employing the corresponding realized measures built from higher-frequency data. The modelling approach features sophisticated asymmetric dynamics in covariances coupled with straightforward estimation and forecasting schemes, independent of the cross-sectional dimension of the assets under consideration. Empirical analyses suggest the HD DCC-HEAVY models have a better in-sample fit and deliver statistically and economically significant out-of-sample gains relative to the standard benchmarks and existing hierarchical factor models. The results are robust under different market conditions.

ABSTRACT. We show that the limiting distribution of HAR test statistics under fixed-b asymptotics is not pivotal when the data are nonstationarity. It takes the form of a complicated function of Gaussian processes and depends on the second moments of the relevant series (e.g., of the regressors and errors for the case of the linear regression model). Hence, existing fixed-b inference methods based on stationarity are not theoretically valid in general. The nuisance parameters entering the fixed-b limiting distribution can be consistently estimated under small-b asymptotics but only with nonparametric rate of convergence. We show that the error in rejection probability (ERP) is an order of magnitude larger than that under stationarity and is also larger than that of HAR tests based on HAC estimators under conventional asymptotics. These theoretical results reconcile with recent finite-sample evidence showing that fixed-b HAR tests can perform poorly when the data are nonstationary. They can be conservative under the null hypothesis and have non-monotonic power under the alternative hypothesis irrespective of how large the sample size is.

ABSTRACT. In this paper, we analyse the computational advantages of the spherical parametri-sation for correlation matrices in the context of Maximum Likelihood estimation via numerical optimisation. By using the special structure of correlation matrices,it is possible to defne a bijective transformation of an n × n correlation matrix Rinto a vector of n(n − 1)∕2 angles between 0 and π. After discussing the algebraic aspects of the problem, we provide examples of the use of the technique we pro-pose in popular econometric models: the multivariate DCC-GARCH model, widely used in applied fnance for large-scale problems, and the multivariate probit model,for which the computation of the likelihood is typically accomplished by simulated Maximum Likelihood. Our analysis reveals the conditions when the spherical par-ametrisation is advantageous; numerical optimisation algorithms are often more robust and efcient, especially when R is large and near-singular.

ABSTRACT. The Misspecification-Resistant Information Criterion (MRIC) proposed in H.-L. Hsu, C.-K. Ing, H. Tong: \emph{On model selection from a finite family of possibly misspecified time series models}. The Annals of Statistics. 47 (2), 1061--1087 (2019), is a model selection criterion for univariate parametric time series that enjoys both the property of consistency and asymptotic efficiency. Its appealing properties make it an ideal tool for time series model selection but, to date, only the univariate response case has been studied. In this article we extend the MRIC to the multivariate time series case. We obtain an asymptotic expression for the mean squared prediction error matrix, we define the vectorial MRIC, and prove the consistency of its method-of-moments estimator. Moreover, we prove its asymptotic efficiency. We discuss the conditions of applicability of the vectorial MRIC for possibly misspecified vector autoregressive models with exogenous variables (VARX) and present a fully worked out example that highlights the need to provide a model selection criterion for multivariate time series that accounts for misspecification.

ABSTRACT. This paper aims to shed light on the effect of the EU ETS on CO2 emissions with a semiparametric panel data model where cross-sectional dependence (CSD) arising from a multifactor error structure and spatial dependence along with heteroscedasticity are all allowed simultaneously. A new estimator that extends the common correlated effect (CCE) approach of Pesaran (2006a) to this framework is proposed. However, the initial estimator ignores the CSD and heteroscedasticity, which will lead to a loss of efficiency. Thus Generalized Least Squares (GLS)-type estimators are proposed. Under rather standard conditions, the parametric estimators are shown to be square root NT-consistent and the asymptotic normality of the nonparametric estimators is also established. Further, the GLS-type estimators are shown to dominate the others. Small sample properties of the estimators are investigated by Monte Carlo experiments and finally by considering a semiparametric extension of the ECK/STIRPAT model, we apply the proposed approach to the dataset consisting of the EU27 countries plus UK and Iceland, and find evidence that the common carbon price arising from the EU ETS policy has a negative nonlinear effect on CO2 emissions.

ABSTRACT. We quantify the relative contributions of the different forces driving climate fluctuations over the past hundreds thousands years. For this aim we estimate historical decomposition of different climate-related time series variables in a structural vector autoregressive analysis in which we allow the number of structural shocks to be smaller than the number of regression residuals. We also exploit non-Gaussianity of the residuals to identify the model, in line with independent component analysis. The results indicate that factors beyond orbital motions have played a significant and relevant role in influencing past climate fluctuations.

ABSTRACT. Climate change has long-run effects on the size and composition of a country’s corporate sector. Using administrative data on the universe of Italian firms, we find that an increase in the incidence of very hot days over a multiyear period persistently reduces the growth rate of the number of active firms in the market. This is due to a drop in firm entry and an increase in firm exit, with relocation playing a minor role. A firm-level investigation reveals a dichotomy between smaller firms, which suffer from high temperatures, and larger firms that successfully adapt, increasing production and net revenues. According to an average climatic scenario, the projected evolution of local temperatures will impact corporate demography further, also exacerbating the divergent effects across warmer and colder areas over the current decade.