nep-for New Economics Papers
on Forecasting
Issue of 2022‒08‒22
six papers chosen by
Rob J Hyndman
Monash University

  1. Superior Predictability of American Factors of the Won/Dollar Real Exchange Rate By Sarthak Behera; Hyeongwoo Kim; Soohyon Kim
  2. A multivariate semi-parametric portfolio risk optimization and forecasting framework By Giuseppe Storti; Chao Wang
  3. LASSO Principal Component Averaging -- a fully automated approach for point forecast pooling By Bartosz Uniejewski; Katarzyna Maciejowska
  4. Forecasting Domestic Tourism across Regional Destinations through MIDAS Regressions. By Nava, Consuelo R.; Osti, Linda; Zoia, Maria Grazia
  5. The Virtue of Complexity Everywhere By Bryan T. Kelly; Semyon Malamud; Kangying Zhou
  6. Nonparametric prediction with spatial data By Gupta, Abhimanyu; Hidalgo, Javier

  1. By: Sarthak Behera; Hyeongwoo Kim; Soohyon Kim
    Abstract: Utilizing an array of data dimensionality reduction methods, we estimate latent common factors for the Won/Dollar real exchange rate from a large panel of economic predictors of the U.S. and Korea. We demonstrate superior out-of-sample predictability of our factor augmented forecasting models relative to conventional models when we utilize factors obtained from U.S. economic variables, while the Korean factors fail to enhance predictability. Our models perform better at longer horizons when the American real activity factors are employed, whereas the American nominal/financial market factors help improve short-run prediction accuracy. The UIP-based global factors with the dollar as numéraire overall perform well, while the PPP and RIRP factors play a limited role in forecasting the Won/Dollar exchange rate.
    Keywords: Won/Dollar Real Exchange Rate; Principal Component Analysis; Partial Least Squares; LASSO; Out-of-Sample Forecast
    JEL: C38 C53 C55 F31 G17
    Date: 2022–07
  2. By: Giuseppe Storti; Chao Wang
    Abstract: A new multivariate semi-parametric risk forecasting framework is proposed, to enable the portfolio Value-at-Risk (VaR) and Expected Shortfall (ES) optimization and forecasting. The proposed framework accounts for the dependence structure among asset returns, without assuming the distribution of returns. A simulation study is conducted to evaluate the finite sample properties of the employed estimator for the proposed model. An empirically motivated portfolio optimization method, that can be utilized to optimize the portfolio VaR and ES, is developed. A forecasting study on 2.5% level evaluates the performance of the model in risk forecasting and portfolio optimization, based on the components of the Dow Jones index for the out-of-sample period from December 2016 to September 2021. Comparing to the standard models in the literature, the empirical results are favourable for the proposed model class, in particular the effectiveness of the proposed framework in portfolio risk optimization is clearly demonstrated.
    Date: 2022–07
  3. By: Bartosz Uniejewski; Katarzyna Maciejowska
    Abstract: This paper develops a novel, fully automated forecast averaging scheme, which combines LASSO estimation method with Principal Component Averaging (PCA). LASSO-PCA (LPCA) explores a pool of predictions based on a single model but calibrated to windows of different sizes. It uses information criteria to select tuning parameters and hence reduces the impact of researchers' at hock decisions. The method is applied to average predictions of hourly day-ahead electricity prices over 650 point forecasts obtained with various lengths of calibration windows. It is evaluated on four European and American markets with almost two and a half year of out-of-sample period and compared to other semi- and fully automated methods, such as simple mean, AW/WAW, LASSO and PCA. The results indicate that the LASSO averaging is very efficient in terms of forecast error reduction, whereas PCA method is robust to the selection of the specification parameter. LPCA inherits the advantages of both methods and outperforms other approaches in terms of MAE, remaining insensitive the the choice of a tuning parameter.
    Date: 2022–07
  4. By: Nava, Consuelo R.; Osti, Linda; Zoia, Maria Grazia (University of Turin)
    Abstract: Over the years, benefits of domestic tourism have been shadowed by the exponential growth of international tourism, despite the former representing a crucial resource, especially at times of geopolitical instability and pandemics. Therefore, forecasting domestic tourism across different regions and sub-regions becomes fundamental to determine its viability as a substitution of international tourism during the COVID-19 pandemic and to evaluate the effectiveness of governmental incentive policies introduced for its promotion. To this aim, and given the availability of data sampled at different frequencies, mixed data-sampling (MIDAS) models have been employed to estimate and predict domestic tourism expenditures, arrivals, and overnight stays. To this aim, we consider the specific case of Italy for illustrative purposes.
    Date: 2022–07
  5. By: Bryan T. Kelly (Yale SOM; AQR Capital Management, LLC; National Bureau of Economic Research (NBER)); Semyon Malamud (Ecole Polytechnique Federale de Lausanne; Centre for Economic Policy Research (CEPR); Swiss Finance Institute); Kangying Zhou (Yale School of Management)
    Abstract: We investigate the performance of non-linear return prediction models in the high complexity regime, i.e., when the number of model parameters exceeds the number of observations. We document a "virtue of complexity" in all asset classes that we study (US equities, international equities, bonds, commodities, currencies, and interest rates). Specifically, return prediction R2 and optimal portfolio Sharpe ratio generally increase with model parameterization for every asset class. The virtue of complexity is present even in extremely data-scarce environments, e.g., for predictive models with less than twenty observations and tens of thousands of predictors. The empirical association between model complexity and out-of-sample model performance exhibits a striking consistency with theoretical predictions.
    Keywords: Portfolio choice, machine learning, random matrix theory, benign overfit, overparameterization
    JEL: C3 C58 C61 G11 G12 G14
    Date: 2022–07
  6. By: Gupta, Abhimanyu; Hidalgo, Javier
    Abstract: We describe a (nonparametric) prediction algorithm for spatial data, based on a canonical factorization of the spectral density function. We provide theoretical results showing that the predictor has desirable asymptotic properties. Finite sample performance is assessed in a Monte Carlo study that also compares our algorithm to a rival nonparametric method based on the infinite AR representation of the dynamics of the data. Finally, we apply our methodology to predict house prices in Los Angeles.
    Keywords: STICERD; ES/R006032/1
    JEL: J1
    Date: 2022–05–23

This nep-for issue is ©2022 by Rob J Hyndman. It is provided as is without any express or implied warranty. It may be freely redistributed in whole or in part for any purpose. If distributed in part, please include this notice.
General information on the NEP project can be found at For comments please write to the director of NEP, Marco Novarese at <>. Put “NEP” in the subject, otherwise your mail may be rejected.
NEP’s infrastructure is sponsored by the School of Economics and Finance of Massey University in New Zealand.