|
on Forecasting |
| By: | Mohamed CHIKHI; Claude DIEBOLT; Tapas MISHRA |
| Abstract: | Despite an inherent share of unpredictability, asset prices such as in stock and Bitcoin markets are naturally driven by significant magnitudes of memory; depending on the strength of path dependence, prices in such markets can be (at least partially) predicted. Being able to predict asset prices is always a boon for investors, more so, if the forecasts are largely unconditional and can only be explained by the series’ own historical trajectories. Although memory dynamics have been exploited in forecasting stock prices, Bitcoin market pose additional challenge, because the lack of proper financial theoretic model limits the development of adequate theory-driven empirical construct. In this paper, we propose a class of autoregressive fractionally integrated moving average (ARFIMA) model with asymmetric exponential generalized autoregressive score (AEGAS) errors to accommodate a complex interplay of ‘memory’ to drive predictive performance (an out-of-sample forecasting). Our conditional variance includes leverage effects, jumps and fat tail-skewness distribution, each of which affects magnitude of memory both the stock and Bitcoin price system would possess enabling us to build a true forecast function. We estimate several models using the Skewed Student-t maximum likelihood and find that the informational shocks in asset prices, in general, have permanent effects on returns. The ARFIMA-AEGAS is appropriate for capturing volatility clustering for both negative (long Value-at-Risk) and positive returns (short Value-at-Risk). We show that this model has better predictive performance over competing models for both long and/or some short time horizons. The predictions from this model beats comfortably the random walk model. Accordingly, we find that the weak efficiency assumption of financial markets stands violated for all price returns studied over longer time horizon. |
| Keywords: | Asset price; Forecasting; Memory; ARFIMA-AEGAS; Leverage effects and jumps; Market Efficiency. |
| JEL: | C14 C58 C22 G17 |
| Date: | 2019 |
| URL: | http://d.repec.org/n?u=RePEc:ulp:sbbeta:2019-43&r=all |
| By: | Yoosoon Chang (Department of Economics, Indiana University); Yongok Choi (School of Economics, Chung-Ang University); Chang Sik Kim (Department of Economics, Sungkyunkwan University); J. Isaac Miller (Department of Economics, University of Missouri-Columbia); Joon Y. Park (Department of Economics, Indiana University and Sungkyunkwan University) |
| Abstract: | Previous authors have pointed out that energy consumption changes both over time and nonlinearly with income level. Recent methodological advances using functional coefficients allow panel models to capture these features succinctly. In order to forecast a functional coefficient out-of-sample, we use functional principal components analysis (FPCA), reducing the problem of forecasting a surface to a much easier problem of forecasting a small number of smoothly varying time series. Using a panel of 180 countries with data since 1971, we forecast energy consumption to 2035 for Germany, Italy, the US, Brazil, China, and India. |
| Keywords: | functional coefficient panel model, functional principal component analysis, energy consumption |
| JEL: | C14 C23 C51 Q43 |
| Date: | 2019–11–25 |
| URL: | http://d.repec.org/n?u=RePEc:umc:wpaper:1915&r=all |
| By: | Bartosz Uniejewski; Rafal Weron |
| Abstract: | Quantile Regression Averaging (QRA) has sparked interest in the electricity price forecasting community after its unprecedented success in the Global Energy Forecasting Competition 2014, where the top two winning teams in the price track used variants of QRA. However, recent studies have reported the method's vulnerability to low quality predictors when the set of regressors is larger than just a few. To address this issue, we consider a regularized variant of QRA, which utilizes the Least Absolute Shrinkage and Selection Operator (LASSO) to automatically select the relevant regressors. We evaluate the introduced technique – dubbed LASSO QRA or LQRA for short – using datasets from the Polish and Nordic power markets, a set of 25 point forecasts obtained for calibration windows of different lengths and 20 different values of the regularization parameter. By comparing against nearly 30 benchmarks, we provide evidence for its superior predictive performance in terms of the Kupiec test, the pinball score and the test for conditional predictive accuracy. |
| Keywords: | Electricity price forecasting; Probabilistic forecast; Quantile Regression Averaging; LASSO; Kupiec test; Pinball score; Conditional predictive accuracy |
| JEL: | C22 C32 C51 C52 C53 Q41 Q47 |
| Date: | 2019–11–16 |
| URL: | http://d.repec.org/n?u=RePEc:wuu:wpaper:hsc1904&r=all |
| By: | Steven F. Lehrer; Tian Xie; Tao Zeng |
| Abstract: | Social media data presents challenges for forecasters since one must convert text into data and deal with issues related to these measures being collected at different frequencies and volumes than traditional financial data. In this paper, we use a deep learning algorithm to measure sentiment within Twitter messages on an hourly basis and introduce a new method to undertake MIDAS that allows for a weaker discounting of historical data that is well-suited for this new data source. To evaluate the performance of approach relative to alternative MIDAS strategies, we conduct an out of sample forecasting exercise for the consumer confidence index with both traditional econometric strategies and machine learning algorithms. Irrespective of the estimator used to conduct forecasts, our results show that (i) including consumer sentiment measures from Twitter greatly improves forecast accuracy, and (ii) there are substantial gains from our proposed MIDAS procedure relative to common alternatives. |
| JEL: | C58 G17 |
| Date: | 2019–11 |
| URL: | http://d.repec.org/n?u=RePEc:nbr:nberwo:26505&r=all |
| By: | Laura Liu (Indiana University, Bloomington, Indiana); Hyungsik Roger Moon (University of Southern California and Yonsei); Frank Schorfheide (University of Pennsylvania CEPR, NBER, and PIER) |
| Abstract: | We use a dynamic panel Tobit model with heteroskedasticity to generate point, set, and density forecasts for a large cross-section of short time series of censored observations. Our fully Bayesian approach allows us to flexibly estimate the cross-sectional distribution of heterogeneous coeffients and then implicitly use this distribution as prior to construct Bayes forecasts for the individual time series. We construct set forecasts that explicitly target the average coverage probability for the cross-section. We present a novel application in which we forecast bank-level charge-off rates for credit card and residential real estate loans, comparing various versions of the panel Tobit model. |
| Keywords: | Bayesian inference, density forecasts, interval forecasts, loan charge-offs, panel data, point forecasts, set forecasts, Tobit model |
| JEL: | C11 C14 C23 C53 G21 |
| Date: | 2019–05 |
| URL: | http://d.repec.org/n?u=RePEc:inu:caeprp:2019005&r=all |
| By: | Dinghai Xu (Department of Economics, University of Waterloo) |
| Abstract: | This paper investigates the “spurious almost integration” effect of volatility under a threshold GARCH structure with realized volatility measures. To closely examine the effect, the realized persistence of volatility is proposed to be used as a threshold trigger for volatility regimes. Under the threshold framework, general closed-form solutions of moment conditions are derived, which provide a convenient way to theoretically examine the “spurious almost integration” effect and its associated impacts. We find that introducing the volatility persistence-driven threshold can capture regime-specific characteristics well. It performs better than the traditional GARCH-type models in terms of both in-sample fitting and out-of-sample forecasting. Based on our Monte Carlo and empirical results, in general we find that overlooking the relatively low persistence regime(s) could lead to some misleading conclusions. |
| JEL: | C01 C58 |
| Date: | 2019–12 |
| URL: | http://d.repec.org/n?u=RePEc:wat:wpaper:1903&r=all |
| By: | Alam, Md. Mahmudul (Universiti Utara Malaysia); Alam, Kazi Ashraful; Shuvo, Anisuzzaman |
| Abstract: | This paper is an attempt to examine the empirical evidence of International Fisher Effect (IFE) between Bangladesh and its two other major trading partners, China and India. The IFE uses interest rate differentials to explain why exchange rates change over time. A time series approach is considered to trace the relationship between nominal interest rates and exchange rates in these countries. The estimated value, by applying OLS, is used to determine the casual relationship between interest rates and exchange rates for quarterly data from 4th Quarter, 1995 to the 2nd Quarter, 2008. The empirical results suggest that there is a little correlation between exchange rates and interest rates differential for Bangladesh with China and Bangladesh with India, and the relationship between the variables is also not noteworthy for Bangladesh. Further, the trends advocate that the forecasting of exchange rates with the hypothesis of IFE is not realistic for these countries. |
| Date: | 2019–02–23 |
| URL: | http://d.repec.org/n?u=RePEc:osf:osfxxx:un95z&r=all |