This paper compares alternative models of time-varying volatility on the basis accuracy real-time point and density forecasts key macroeconomic time series for USA. We consider Bayesian autoregressive vector that incorporate some form volatility, precisely random walk stochastic following a stationary AR process, coupled with fat tails, GARCH mixture innovation models. The results show VAR specifications conventional dominate other specifications, in terms forecasting to degree greater...

This article revisits the accuracy of inflation forecasting using activity and expectations variables. We apply Bayesian model averaging across different regression specifications selected from a set potential predictors that includes lagged values inflation, host real data, term structure (relative) price surveys. In this average, we can entertain channels structural instability, by either incorporating stochastic breaks in parameters each individual specification within or allowing for...

Abstract In this paper, we empirically evaluate competing approaches for combining inflation density forecasts in terms of Kullback–Leibler divergence. particular, apply a similar suite models to four different datasets and aim at identifying combination methods that perform well throughout series variations the model suite. We pool individual densities using linear logarithmic methods. The consists forecasting with moving estimation windows account structural change. find is much better...

In public discussions of the quality forecasts, attention typically focuses on predictive performance in cases extreme events. However, restriction conventional forecast evaluation methods to subsets observations has unexpected and undesired effects, is bound discredit skillful forecasts when signal-to-noise ratio data generating process low. Conditioning outcomes incompatible with theoretical assumptions established methods, thereby confronting forecasters what we refer as forecaster's...

We estimate demand, supply, monetary, investment and financial shocks in a VAR identified with minimum set of sign restrictions on US data. find that are major drivers fluctuations output, stock prices but have limited effect inflation. In second step, we disentangle originating the housing sector, credit markets uncertainty shocks. extended set‐up, even more important leading role is played by large persistent effects output.

One of the most controversial issues in mid-term load forecasting literature is treatment weather. Because difficulty obtaining precise weather forecasts for a few weeks ahead, researchers have, so far, implemented three approaches: a) excluding from models altogether, b) assuming future to be perfectly known and c) including their models. This article provides first systematic comparison how different treatments affect performance. We incorporate air temperature into short- models,...

This study examines the impact of COVID-19 pandemic on corporate financial performance using a unique, cross-country, and longitudinal sample 3350 listed firms worldwide. We find that family has been significantly higher than nonfamily during pandemic, accounting for pre-pandemic business conditions. effect is pertinent to with strong involvement in management or both ownership. also identify role firm-, industry-, country-level contingencies pandemic. offers novel understanding resilience...

We study the real‐time predictive content of crude oil prices for U.S. real GDP growth through a pseudo out‐of‐sample (OOS) forecasting exercise. Comparing our benchmark model “without oil” against alternatives “with oil,” we strongly reject null hypothesis no OOS population‐level predictability from to at longer forecast horizon consider. This examination global relative performance models consider is robust use ex post revised data. But when focus on models’ local performance, observe...

Summary We propose a density combination approach featuring weights that depend on the past forecast performance of individual models entering through utility‐based objective function. apply this model scheme to stock returns, both at aggregate level and by industry, investigate its forecasting relative host existing methods, within class linear time‐varying coefficients, stochastic volatility models. Overall, we find our produces markedly more accurate predictions than alternatives, in...

We introduce a combined density nowcasting (CDN) approach to dynamic factor models (DFM) that in coherent way accounts for time-varying uncertainty of several model and data features provide more accurate complete nowcasts. The combination weights are latent random variables depend on past performance other learning mechanisms. scheme is incorporated Bayesian sequential Monte Carlo method which rebalances the set nowcasted densities each period using updated information weights. Experiments...

The ongoing transformation of the electricity market has reshaped hydropower production paradigm for storage reservoir systems, with a shift from strategies oriented towards maximizing regional energy to aimed at revenue maximization individual systems. Indeed, producers bid their scheduling 1 day in advance, attempting align operational plan hours where expected prices are higher. As result, accuracy 1-day ahead forecasts started play key role short-term optimization This paper aims...

This paper analyzes sovereign risk shift-contagion, i.e. positive and significant changes in the propagation mechanisms, using bond yield spreads for major eurozone countries. By emphasizing use of two econometric approaches based on quantile regressions (standard regression Bayesian with heteroskedasticity) we find that shocks euro's shows almost no presence shift-contagion. All increases correlation have witnessed over last years come from larger propagated higher intensity across Europe.

We introduce a Bayesian approach to predictive density calibration and combination that accounts for parameter uncertainty model set incompleteness through the use of random functionals weights. Building on work Ranjan Gneiting, we infinite beta mixtures calibration. The proposed nonparametric takes advantage flexibility Dirichlet process achieve any continuous deformation linearly combined distributions. inference procedure is based Gibbs slice sampling. provide some conditions under which...

Summary The proposed panel Markov‐switching VAR model accommodates changes in low and high data frequencies incorporates endogenous time‐varying transition matrices of country‐specific Markov chains, allowing for interconnections. An efficient multi‐move sampling algorithm draws chains. Using industrial production growth credit spread data, several important features are obtained. Three regimes appear, with slow becoming persistent the eurozone. Turning point analysis indicates USA leading...

We compare alternative univariate versus multivariate models and frequentist Bayesian autoregressive vector specifications for hourly day-ahead electricity prices, both with without renewable energy sources. The accuracy of point density forecasts is inspected in four main European markets (Germany, Denmark, Italy, Spain) characterized by different levels power generation. Our results show that the exogenous variables dominate other terms forecasting forecasting.

This article shows entropic tilting to be a flexible and powerful tool for combining medium-term forecasts from BVARs with short-term other sources (nowcasts either surveys or models). Tilting systematically improves the accuracy of both point density forecasts, BVAR based on nowcast means variances yields slightly greater gains in than does just means. Hence, can offer—more so persistent variables not-persistent variables—some benefits accurately estimating uncertainty multi-step that...

Abstract This paper proposes world steel production as an indicator of global real economic activity. World data is published with only a one‐month delay, thereby providing timely information for GDP forecasters. We find that and Lutz Kilian's (2009) index activity generate large gains in forecasting GDP, relative to autoregressive benchmark. A forecast combination production, the industrial OECD countries plus six non‐OECD emerging economies produces significant benchmark