|
on Forecasting |
| By: | Oscar Claveria (Department of Econometrics. University of Barcelona ); Enric Monte (Department of Signal Theory and Communications. Polytechnic University of Catalunya. ); Salvador Torra (Department of Econometrics & Riskcenter-IREA. Universitat de Barcelona ) |
| Abstract: | This study aims to analyze the effects of data pre-processing on the performance of forecasting based on neural network models. We use three different Artificial Neural Networks techniques to forecast tourist demand: a multi-layer perceptron, a radial basis function and an Elman neural network. The structure of the networks is based on a multiple-input multiple-output setting (i.e. all countries are forecasted simultaneously). We use official statistical data of inbound international tourism demand to Catalonia (Spain) and compare the forecasting accuracy of four processing methods for the input vector of the networks: levels, growth rates, seasonally adjusted levels and seasonally adjusted growth rates. When comparing the forecasting accuracy of the different inputs for each visitor market and for different forecasting horizons, we obtain significantly better forecasts with levels than with growth rates. We also find that seasonally adjusted series significantly improve the forecasting performance of the networks, which hints at the significance of deseasonalizing the time series when using neural networks with forecasting purposes. These results reveal that, when using seasonal data, neural networks performance can be significantly improved by working directly with seasonally adjusted levels. |
| Keywords: | Artificial neural networks, forecasting, multiple-input multiple-output (MIMO), seasonality, detrending, tourism demand, multilayer perceptron, radial basis function, Elman JEL classification: L83, C53, C45, R11 |
| Date: | 2015–01 |
| URL: | http://d.repec.org/n?u=RePEc:aqr:wpaper:201503&r=for |
| By: | Chin, Michael (Bank of England ); Polk, Christopher (London School of Economics ) |
| Abstract: | Recent studies find evidence in favour of return predictability, and argue that their positive findings result from their ability to capture expected returns. We assess the forecasting performance of two popular approaches to estimating expected equity returns, a dividend discount model (DDM) commonly used to estimate `implied cost of capital', and a vector autoregression (VAR) model commonly used to decompose equity returns. In line with recent evidence, in-sample tests show that both estimates generate substantially lower forecast errors compared to traditional predictor variables such as price-earnings ratios and dividend yields. Out-of-sample, the VAR and DDM estimates generate economically and statistically significant forecast improvements relative to a historical average benchmark. Our results tentatively suggest that the VAR approach better captures expected returns compared to the DDM. |
| Keywords: | Expected returns; implied cost of capital; dividend discount model; return predictability; forecasting |
| JEL: | G10 G11 G12 G17 |
| Date: | 2015–01–16 |
| URL: | http://d.repec.org/n?u=RePEc:boe:boeewp:0520&r=for |
| By: | Sepideh Dolatabadi (Queen's University ); Paresh Kumar Narayan (Deakin University ); Morten Ørregaard Nielsen (Queen's University and CREATES ); Ke Xu (Queen's University ) |
| Abstract: | Based on recent evidence of fractional cointegration in commodity spot and futures markets, we investigate whether a fractionally cointegrated model can provide statistically and/or economically significant forecasts of commodity returns. Specifically, we propose to model and forecast commodity spot and futures prices using a fractionally cointegrated vector autoregressive model that generalizes the more well-known cointegrated vector autoregressive model to allow fractional integration. We derive the best linear predictor forecast for this model and perform an out-of-sample forecast comparison with forecasts from the more standard (non-fractional) model. In our empirical analysis to daily data on 15 commodity spot and futures markets, the fractional model is found to be superior in terms of in-sample fit and also out-of-sample forecasting based on statistical metrics of forecast comparison. We then analyze the economic significance of the forecasts through a dynamic trading strategy based on a portfolio with weights derived from a mean-variance utility function. This analysis leads to statistically significant and economically meaningful profits in the commodity markets, and also shows that the fractional model generates higher profits on average compared with the non-fractional model. |
| Keywords: | commodity markets, economic significance, forecasting, fractional cointegration, futures markets, price discovery, vector error correction model |
| JEL: | C32 G11 |
| Date: | 2015–01 |
| URL: | http://d.repec.org/n?u=RePEc:qed:wpaper:1337&r=for |
| By: | Knüppel, Malte |
| Abstract: | Recently, several institutions have increased their forecast horizons, and many institutions rely on their past forecast errors to estimate measures of forecast uncertainty. This work addresses the question how the latter estimation can be accomplished if there are only very few errors available for the new forecast horizons. It extends upon the results of Knüppel (2014) in order to relax the condition on the data structure required for the SUR estimator to be independent from unknown quantities. It turns out that the SUR estimator of forecast uncertainty tends to deliver large e¢ ciency gains compared to the OLS estimator (i.e. the sample mean of the squared forecast errors) in the case of increased forecast horizons. The SUR estimator is applied to the forecast errors of the Bank of England and the FOMC. |
| Keywords: | multi-step-ahead forecasts,forecast error variance,SUR |
| JEL: | C13 C32 C53 |
| Date: | 2014 |
| URL: | http://d.repec.org/n?u=RePEc:zbw:bubdps:402014&r=for |
| By: | Tim Bollerslev (Duke University, NBER and CREATES ); Andrew J. Patton (Duke University, NBER and CREATES ); Wenjing Wang (Moody’s Analytics, Inc. ) |
| Abstract: | We construct daily house price indices for ten major U.S. metropolitan areas. Our calculations are based on a comprehensive database of several million residential property transactions and a standard repeat-sales method that closely mimics the methodology of the popular monthly Case-Shiller house price indices. Our new daily house price indices exhibit dynamic features similar to those of other daily asset prices, with mild autocorrelation and strong conditional heteroskedasticity of the corresponding daily returns. A relatively simple multivariate time series model for the daily house price index returns, explicitly allowing for commonalities across cities and GARCH effects, produces forecasts of monthly house price changes that are superior to various alternative forecast procedures based on lower frequency data. |
| Keywords: | Data aggregation, Real estate prices, Forecasting, Time-varying volatility |
| JEL: | C22 C32 C53 G17 R21 |
| Date: | 2015–01–12 |
| URL: | http://d.repec.org/n?u=RePEc:aah:create:2015-02&r=for |
| By: | Oscar Claveria (Department of Econometrics. University of Barcelona ); Enric Monte (Department of Signal Theory and Communications. Polytechnic University of Catalunya. ); Salvador Torra (Department of Econometrics & Riskcenter-IREA. Universitat de Barcelona ) |
| Abstract: | This study attempts to improve the forecasting accuracy of tourism demand by using the existing common trends in tourist arrivals form all visitor markets to a specific destination in a multiple-input multiple-output (MIMO) structure. While most tourism forecasting research focuses on univariate methods, we compare the performance of three different Artificial Neural Networks in a multivariate setting that takes into account the correlations in the evolution of inbound international tourism demand to Catalonia (Spain). We find that the MIMO approach does not outperform the forecasting accuracy of the networks when applied country by country, but it significantly improves the forecasting performance for total tourist arrivals. When comparing the forecast accuracy of the different models, we find that radial basis function networks outperform multilayer-perceptron and Elman networks. |
| Keywords: | Tourism demand, forecasting, multivariate, multiple-output, artificial neural networks JEL classification: C22, C45, C63, L83, R11 |
| Date: | 2015–01 |
| URL: | http://d.repec.org/n?u=RePEc:aqr:wpaper:201502&r=for |
| By: | Hossein Hassani (The Statistical Research Centre, Bournemouth University, UK ); Zara Ghodsi (The Statistical Research Centre, Bournemouth University, UK ); Rangan Gupta (Department of Economics, University of Pretoria, Pretoria, South Africa ); Mawuli K. Segnon (Christian-Albrechts-University Kiel, Department of Economics, 24098, Kiel, Germany ) |
| Abstract: | Accurate forecasts of home sales can provide valuable information for not only, policy makers, but also financial institutions and real estate professionals. Given this, our analysis compares the ability of two different versions of Singular Spectrum Analysis (SSA) methods, namely Recurrent SSA (RSSA) and Vector SSA (VSSA), in univariate and multivariate frameworks, in forecasting seasonally unadjusted home sales for the aggregate US economy and its four census regions (Northeast, Midwest, South and West). We compare the performance of the SSA-based models with classical and Bayesian variants of the autoregressive and vector autoregressive models. Using an out-of-sample period of 1979:8-2014:6, given an in-sample period of 1973:1-1979:7, we find that the univariate VSSA is the best performing model for the aggregate US home sales, while the multivariate versions of the RSSA is the outright favorite in forecasting home sales for all the four census regions. Our results highlight the superiority of the nonparametric approach of the SSA, which in turn, allows us to handle any statistical process: linear or nonlinear, stationary or non-stationary, Gaussian or non-Gaussian. |
| Keywords: | Home Sales, Forecasting, Singular Spectrum Analysis, Classical and Bayesian (Vector) Autoregressive Models |
| JEL: | C32 R31 |
| Date: | 2014–12 |
| URL: | http://d.repec.org/n?u=RePEc:pre:wpaper:201482&r=for |
| By: | Recchioni, Maria Cristina ; Tedeschi, Gabriele ; Gallegati, Mauro |
| Abstract: | In this paper we introduce a calibration procedure for validating of agent based models. Starting from the well-known financial model of Brock and Hommes 1998, we show how an appropriate calibration enables the model to describe price time series. We formulate the calibration problem as a nonlinear constrained optimization that can be solved numerically via a gradient-based method. The calibration results show that the simplest version of the Brock and Hommes model, with two trader types, fundamentalists and trend-followers, replicates nicely the price series of four different markets indices: the S&P 500, the Euro Stoxx 50, the Nikkei 225 and the CSI 300. We show how the parameter values of the calibrated model are important in interpret- ing the trader behavior in the different markets investigated. These parameters are then used for price forecasting. To further improve the forecasting, we modify our calibration approach by increasing the trader information set. Finally, we show how this new approach improves the model's ability to predict market prices. |
| Keywords: | Calibration,Validation,Forecasting,Agent-based models,Asset pricing,Heterogeneous beliefs |
| JEL: | C53 C63 G17 |
| Date: | 2014 |
| URL: | http://d.repec.org/n?u=RePEc:zbw:fmpwps:26&r=for |
| By: | Magdalena Szyszko (Wyzsza Szkola Bankowa w Poznaniu, Poland ); Karolina Tura (Uniwersytet Ekonomiczny w Poznaniu, Poland ) |
| Abstract: | Producing and revealing inflation forecasts is believed to be the best way of implementing a forward-looking monetary policy. The article focuses on inflation forecast targeting (IFT) at the Czech National Bank (CNB) in terms of its efficiency in shaping consumers’ inflation expectations. The goal of the study is to verify accuracy of the inflation forecasts, and their influence on inflation expectations. The research is divided into four stages. At the first stage central bank credibility is examined. At the second stage – accuracy of the inflation forecasts. The next step covers a qualitative analysis of IFT implementation. Finally the existence of the interdependences of inflation forecast, optimal policy paths and inflation expectations are analyzed. Credibility of the central bank, accuracy of the forecast and decision-making procedures are the premises for the existence of relationship between forecasts and expectations. The research covers July 2002 - end of 2013. Its methodology includes the qualitative analysis of decision-making of the CNB, quantitative methods (Kia and Patron formula, MAE forecasts errors, quantification of expectations, non-parametric statistics). The results show the existence of interdependences between inflation forecasts and expectations of moderate strength. The preconditions of such interdependences are partially fulfilled. The research opens the field for cross-country comparisons and for quantification of IFT implementation. |
| Keywords: | inflation forecasts, inflation forecast targeting, policy path, inflation expectations |
| JEL: | E52 E58 E61 |
| Date: | 2014–12 |
| URL: | http://d.repec.org/n?u=RePEc:pes:wpaper:2014:no49&r=for |
| By: | Domenico Ferraro ; Ken Rogoff ; Barbara Rossi |
| Abstract: | We show the existence of a very short-term relationship at the daily frequency between changes in the price of a country's major commodity export price and changes in its nominal exchange rate. The relationship appears to be robust and to hold when we use contemporaneous (realized) commodity price changes in our regression. However, when we use lagged commodity price changes, the predictive ability is ephemeral, mostly appearing after instabilities have been appropriately taken into account. |
| JEL: | F31 F37 C22 C53 |
| Date: | 2011–05 |
| URL: | http://d.repec.org/n?u=RePEc:upf:upfgen:1461&r=for |