|
on Forecasting |
| By: | Degiannakis, Stavros; Filis, George |
| Abstract: | Contrary to the current practice that mainly considers stand-alone statistical loss functions, the aim of the paper is to assess oil price volatility forecasts based on objective-based evaluation criteria, given that different forecasting models may exhibit superior performance at different applications. To do so, we forecast implied and several intraday volatilities and we evaluate them based on financial decisions for which these forecasts are used. In this study we confine our interest on the use of such forecasts from financial investors. More specifically, we consider four well established trading strategies, which are based on volatility forecasts, namely (i) trading the implied volatility based on the implied volatility forecasts, (ii) trading implied volatility based on intraday volatility forecasts, (iii) trading straddles in the United States Oil Fund ETF and finally (iv) trading the United States Oil Fund ETF based on implied and intraday volatility forecasts. We evaluate the after-cost profitability of each forecasting model for 1-day up to 66-days ahead. Our results convincingly show that our forecasting framework is economically useful, since different models provide superior after-cost profits depending on the economic use of the volatility forecasts. Should investors evaluate the forecasting models based on statistical loss functions, then their financial decisions would be sub-optimal. Thus, we maintain that volatility forecasts should be evaluated based on their economic use, rather than statistical loss functions. Several robustness tests confirm these findings. |
| Keywords: | Volatility forecasting, implied volatility, intraday volatility, WTI crude oil futures, objective-based evaluation criteria. |
| JEL: | C22 C53 G11 G17 Q47 |
| Date: | 2019–06–11 |
| URL: | http://d.repec.org/n?u=RePEc:pra:mprapa:94445&r=all |
| By: | Pourghorban, Mojtaba; Mamipour, Siab |
| Abstract: | This paper provides a method to forecast day-ahead electricity prices based on autoregressive integrated moving average (ARIMA) and generalized autoregressive conditional heteroskedastic (GARCH) models. In the competitive power market environment, electricity price forecasting is an essential task for market participants. However, time series of electricity price has complex behavior such as nonlinearity, nonstationarity, and high volatility. ARIMA is suitable in forecasting, but it is not able to handle nonlinearity and volatility are existent in time series. Therefore, GARCH models are used to handle volatility in the in time series forecasting. The proposed method is computed using the daily electricity price data of Iran market for a five-year period from March 2013 to February 2018. The results reported in this paper illustrate the potential of the proposed ARMA-GARCH model and this combined model has been successfully applied to real prices in the Iranian power market. |
| Keywords: | Electricity price forecasting, ARIMA model, GARCH model |
| JEL: | C3 C32 C5 C53 Q4 Q47 |
| Date: | 2019–02–14 |
| URL: | http://d.repec.org/n?u=RePEc:pra:mprapa:94826&r=all |
| By: | Joshua C. C. Chan; Liana Jacobi; Dan Zhu |
| Abstract: | Large Bayesian VARs with the natural conjugate prior are now routinely used for forecasting and structural analysis. It has been shown that selecting the prior hyperparameters in a data-driven manner can often substantially improve forecast performance. We propose a computationally efficient method to obtain the optimal hyperparameters based on Automatic Differentiation, which is an efficient way to compute derivatives. Using a large US dataset, we show that using the optimal hyperparameter values leads to substantially better forecast performance. Moreover, the proposed method is much faster than the conventional grid-search approach, and is applicable in high-dimensional optimization problems. The new method thus provides a practical and systematic way to develop better shrinkage priors for forecasting in a data-rich environment. |
| Keywords: | automatic differentiation, vector autoregression, optimal hyperparameters, forecasts, marginal likelihood |
| JEL: | C11 C53 E37 |
| Date: | 2019–06 |
| URL: | http://d.repec.org/n?u=RePEc:een:camaaa:2019-46&r=all |
| By: | Shuo Cao (Shenzhen Stock Exchange); Huichou Huang (Broad Reach Investment Management); Ruirui Liu (King’s College London); Ronald MacDonald (University of Glasgow) |
| Abstract: | In this paper we study the exchange rate predictability across a range of investment horizons by proposing a generalized (term structure) model to capture the risk premium component of exchange rates with a broad set of variables meanwhile handle both parameter and model uncertainty. We demonstrate the existence of time-varying term-structural effect and model disagreement effect of exchange rate predictors as well as the projections of predictive information over the term structure. We further utilize the time-variation in the probability weighting to identify the scapegoat drivers of customer order flows. Our findings suggest that heterogeneous agents learn to forecast exchange rates and switch trading rules over time, resulting in the dynamic country-specific and global exposures of exchange rates to short- run non-fundamental risk and long-run business cycle risk. Hedging pressure and liquidity are identified to contain predictive information that is common to a range of forecasting horizons. Policy-related predictors are important for short-run forecasts up to 3 months while crash risk indicators matter for long-run forecasts from 9 months to 12 months. We further comprehensively evaluate both statistical and economic significance of the model allowing for a full spectrum of currency investment management, and find that the model generates substantial performance fees of 6.5% per annum. The outperformance is mainly due to (i) the relaxing of restrictions imposed on structural parameters via model generalization, and (ii) the use of factor structure to extract common useful information from noisy data and reduce estimation errors. |
| Keywords: | Exchange Rate Forecasting, Disconnect Puzzle, Carry Trade Risk Premia, Term Structure Factors, Scapegoat Variables, Model Disagreement, Customer Order Flows |
| JEL: | C52 E43 F31 F37 G11 |
| URL: | http://d.repec.org/n?u=RePEc:cth:wpaper:gru_2017_013&r=all |
| By: | Manabu Asai (Faculty of Economics, Soka University, Japan); Rangan Gupta (Department of Economics, University of Pretoria, Pretoria, South Africa); Michael McAleer (Department of Finance, Asia University, Taiwan; Discipline of Business Analytics, University of Sydney Business School, Australia; Econometric Institute, Erasmus School of Economics, Erasmus University Rotterdam, The Netherlands; Department of Economic Analysis and ICAE Complutense University of Madrid, Spain and Institute of Advanced Sciences, Yokohama National University, Japan) |
| Abstract: | For purposes of forecasting the covariance matrix for the returns of crude oil and gold futures, this paper examines the effects of leverage, jumps, spillovers, and geopolitical risks, using their respective realized covariance matrices. In order to guarantee the positive definiteness of the forecasts, we consider the full BEKK structure on the conditional Wishart model. By the specification, we can divide flexibly the direct and spillover effects of volatility feedback, negative returns, and jumps. The empirical analysis indicates the benefits in accommodating the spillover effects of negative returns and the geopolitical risks indicator for modelling and forecasting the future covariance matrix. |
| Keywords: | Commodity Markets, Co-volatility, Forecasting, Geopolitical Risks, Jumps, Leverage Effects, Spillover Effects, Realized Covariance |
| JEL: | C32 C33 C58 Q02 |
| Date: | 2019–07 |
| URL: | http://d.repec.org/n?u=RePEc:pre:wpaper:201951&r=all |
| By: | Anthony Garratt; Timo Henckel; Shaun P. Vahey |
| Abstract: | Many studies have found that combining density forecasts improves predictive accuracy for macroeconomic variables. A prevalent approach known as the Linear Opinion Pool (LOP) combines forecast densities from “experts”; see, among others, Stone (1961), Geweke and Amisano (2011), Kascha and Ravazzolo (2011), Ranjan and Gneiting (2010) and Gneiting and Ranjan (2013). Since the LOP approach averages the experts’ probabilistic assessments, the distribution of the combination generally differs from the marginal distributions of the experts. As a result, the LOP combination forecasts sometimes fail to match salient features of the sample data, including asymmetries in risk. In this paper, we propose a computationally convenient transformation for a target macroeconomic variable with an asymmetric marginal distribution. Our methodology involves a Smirnov transform to reshape the LOP combination forecasts using a nonparametric kernel-smoothed empirical cumulative distribution function. We illustrate our methodology with an application examining quarterly real-time forecasts for US inflation based on multiple output gap measures over an evaluation sample from 1990:1 to 2017:2. Our proposed methodology improves combination forecast performance by approximately 10% in terms of both the root mean squared forecast error and the continuous ranked probability score. We find that our methodology delivers a similar performance gain for the Logarithmic Opinion Pool (LogOP), a commonly-used alternative to the LOP. |
| Keywords: | Forecast density combination, Smirnov transform, inflation |
| Date: | 2019–07 |
| URL: | http://d.repec.org/n?u=RePEc:een:camaaa:2019-47&r=all |
| By: | Kajal Lahiri; Wuwei Wang |
| Abstract: | We apply generalized beta and triangular distributions to histograms from the Survey of Professional Forecasters (SPF) to estimate forecast uncertainty, shocks and discord using information framework, and compare these with moment-based estimates. We find these two approaches to produce analogous results, except in cases where the underlying densities deviate significantly from normality. Even though the Shannon entropy is more inclusive of different facets of a forecast density, we find that with SPF forecasts it is largely driven by the variance of the densities. We use Jenson-Shannon Information to measure ex ante “news” or “uncertainty shocks” in real time, and find that this ‘news’ is closely related to revisions in forecast means, countercyclical, and raises uncertainty. Using standard vector auto-regression analysis, we confirm that uncertainty affects the economy negatively. |
| Keywords: | density forecasts, uncertainty, disagreement, entropy measures, Jensen-Shannon information, Survey of Professional Forecasters |
| JEL: | E37 |
| Date: | 2019 |
| URL: | http://d.repec.org/n?u=RePEc:ces:ceswps:_7674&r=all |