nep-for New Economics Papers
on Forecasting
Issue of 2012‒10‒27
sixteen papers chosen by
Rob J Hyndman
Monash University

  1. Can we beat the random walk in forecasting CEE exchange rates? By Jakub Muck; Pawel Skrzypczynski
  2. Nowcasting German GDP: A comparison of bridge and factor models. By Antipa, P.; Barhoumi, K.; Brunhes-Lesage, V.; Darné, O.
  3. Imputing Individual Effects in Dynamic Microsimulation Models. An application of the Rank Method By Matteo Richiardi; Ambra Poggi
  4. Forecasting 2012 United States Presidential election using Factor Analysis, Logit and Probit Models By Sinha, Pankaj; Thomas, Ashley Rose; Ranjan, Varun
  5. Real exchange rate forecasting: a calibrated half-life PPP model can beat the random walk By Michele Ca’ Zorzi; Michal Rubaszek
  6. Forecasting Chinese inflation and output: A Bayesian vector autoregressive approach By Huang, Y-F.
  7. Non-Parametric Stochastic Simulations to Investigate Uncertainty around the OECD Indicator Model Forecasts By Elena Rusticelli
  8. Bilinear forecast risk assessment for non-systematic inflation: Theory and evidence By Wojciech Charemza; Yuriy Kharin; Vladislav Maevskiy
  9. MPC Voting, Forecasting and Inflation By Wojciech Charemza; Daniel Ladley
  10. Russia's Financial Markets Development Forecast till 2020 By Àlexey Vedev
  11. Prior Selection for Vector Autoregressions By Domenico Giannone; Michele Lenza; Giorgio E. Primiceri
  12. Modelling the impact of aggregate financial shocks external to the Chinese economy By Qin, Duo; He, Xinhua
  13. SEMIFARMA-HYGARCH Modeling of Dow Jones Return Persistence By Mohamed Chikhi; Anne Péguin-Feissolle; Michel Terraza
  14. Probability and Asset Updating using Bayesian Networks for Combinatorial Prediction Markets By Wei Sun; Robin Hanson; Kathryn Blackmond Laskey; Charles Twardy
  15. Can We Forecast the Implied Volatility Surface Dynamics of Equity Options? Predictability and Economic Value Tests By Alejandro Bernales; Massimo Guidolin
  16. A Copula Based Bayesian Approach for Paid-Incurred Claims Models for Non-Life Insurance Reserving By Gareth W. Peters; Alice X. D. Dong; Robert Kohn

  1. By: Jakub Muck (National Bank of Poland, Economic Institute; Warsaw School of Economics, Institute of Econometrics); Pawel Skrzypczynski (National Bank of Poland, Economic Institute)
    Abstract: It is commonly known that various econometric techniques fail to consistently outperform a simple random walk model in forecasting exchange rates. The aim of this study is to analyse whether this also holds for selected currencies of the CEE region as the literature relating to the ability of forecasting these exchange rates is scarce. We tackle this issue by comparing the random walk based out-of-sample forecast errors of the Polish zloty, the Czech koruna and the Hungarian forint exchange rates against the euro with the corresponding errors generated by various single- and multi-equation models of these exchange rates. The results confirm that it is very difficult to outperform a simple random walk model in our CEE currencies forecasting contest.
    Keywords: CEE currencies, exchange rate forecasting, random walk,VAR, BVAR
    JEL: C22 C32 C53 F31 G17
    Date: 2012
  2. By: Antipa, P.; Barhoumi, K.; Brunhes-Lesage, V.; Darné, O.
    Abstract: Governments and central banks need to have an accurate and timely assessment of Gross Domestic Product's (GDP) growth rate for the current quarter, as this is essential for providing a reliable and early analysis of the current economic situation. This paper presents a series of models conceived to forecast the current German GDP's quarterly growth rate. These models are designed to be used on a monthly basis by integrating monthly economic information through bridge models, thus allowing for the economic interpretation of the data. We do also forecast German GDP by dynamic factor models. The combination of these two approaches allows selecting economically relevant explanatory variables among a large data set of hard and soft data. In addition, a rolling forecast study is carried out to assess the forecasting performance of the estimated models. To this end, publication lags are taken into account in order to run pseudo out-of-sample forecasts. We show that it is possible to get reasonably good estimates of current quarterly GDP growth in anticipation of the official release, especially from bridge models.
    Keywords: GDP forecasting; Bridge models; Factor models.Threshold auto-regression, bounce-back effects, business cycles, inventory investment.
    JEL: C52 C53 E20
    Date: 2012
  3. By: Matteo Richiardi; Ambra Poggi
    Abstract: Dynamic microsimulation modeling involves two stages: estimation and forecasting. Unobserved heterogeneity is often considered in estimation, but not in forecasting, beyond trivial cases. Non-trivial cases involve individuals that enter the simulation with a history of previous outcomes. We show that the simple solutions of attributing to these individuals a null effect or a random draw from the estimated unconditional distributions lead to biased forecasts, which are often worse than those obtained neglecting unobserved heterogeneity altogether. We then present a first implementation of the Rank method, a new algorithm for attributing the individual effects to the simulation sample which greatly simplifies those already known in the literature. Out-of-sample validation of our model shows that correctly imputing unobserved heterogeneity significantly improves the quality of the forecasts.
    Keywords: Dynamic microsimulation, Unobserved heterogeneity, Validation, Rank method, Assignment algorithms, Female labor force participation, Italy
    JEL: C53 C18 C23 C25 J11 J12 J21
    Date: 2012
  4. By: Sinha, Pankaj; Thomas, Ashley Rose; Ranjan, Varun
    Abstract: Contemporary discussions on 2012 U.S Presidential election mention that economic variables such as unemployment rate, inflation, budget deficit/surplus, public debt, tax policy and healthcare spending will be deciding elements in the forthcoming November election. Certain researchers like Bartells and Zaller (2001), Lewis-Beck and Rice (1982), and Lichtman and Keilis-Borok (1996) have investigated the significance of non-economic variables in forecasting the U.S election. This paper investigates the influence of combination of various economic and non-economic variables as factors influencing the outcome of 2012 U.S Presidential election, using statistical factor analysis. The obtained factor scores are used to predict the vote share of the incumbent using regression model. The paper also employs logit and probit models to predict the probability of win for the incumbent candidate in 2012 U.S Presidential election. It is found that the factors combining above economic variables are insignificant in deciding the outcome of the 2012 election. The factor combining the non-economic variables such as Gallup Ratings, GIndex, wars and scandals has been found significantly influencing the public perception of the performance of the Government and its policies, which in turn affects the voting decision. The proposed factor regression model forecasts that the Democrat candidate Mr. Barack Obama is likely to get a vote share between 51.84% - 54.26% with 95% confidence interval in the forthcoming November 2012 U.S Presidential election. While, the proposed logit and probit models forecast the probability of win for the Democrat candidate Mr. Barack Obama to be 67.37% and 67.00%, respectively.
    Keywords: Factor Analysis; Logit and Probit model; 2012 U.S Presidential Election; Economic and non-economic variables
    JEL: C53 C5 D72 C2 C01 C1
    Date: 2012–10–15
  5. By: Michele Ca’ Zorzi (European Central Bank); Michal Rubaszek (National Bank of Poland, Economic Institute; Warsaw School of Economics)
    Abstract: This paper brings two new insights into the Purchasing Power Parity (PPP) debate. First, even if PPP is thought to hold only in the long run, we show that a half-life PPP model outperforms the random walk in real exchange rate forecasting, also at short-term horizons. Second, we show that this result holds as long as the speed of adjustment to the sample mean is imposed and not estimated. The reason is that the estimation error of the pace of convergence distorts the results in favor of the random walk model, even if the PPP holds in the long-run.
    Keywords: Exchange rate forecasting; purchasing power parity; half-life
    JEL: C32 F31 F37
    Date: 2012
  6. By: Huang, Y-F.
    Abstract: This study compares several Bayesian vector autoregressive (VAR) models for forecasting price inflation and output growth in China. The results indicate that models with shrinkage and model selection priors, that restrict some VAR coefficients to be close to zero, perform better than models with Normal prior.
    Keywords: BVAR; factor model; shrinkage priors
    JEL: C32 C11
    Date: 2012–10
  7. By: Elena Rusticelli
    Abstract: The forecasting uncertainty around point macroeconomic forecasts is usually measured by the historical performance of the forecasting model, using measures such as root mean squared forecasting errors (RMSE). This measure, however, has the major drawback that it is constant over time and hence does not convey any information on the specific source of uncertainty nor the magnitude and balance of risks in the immediate conjuncture. Moreover, specific parametric assumptions on the probability distribution of forecasting errors are needed in order to draw confidence bands around point forecasts. This paper proposes an alternative time-varying simulated RMSE, obtained by means of non-parametric stochastic simulations, which combines the uncertainty around the model’s parameters and the structural errors term to construct asymmetric confidence bands around point forecasts. The procedure is applied, by way of example, to the short-term real GDP growth forecasts generated by the OECD Indicator Model for Germany. The empirical probability distributions of the GDP growth forecasts, derived through the bootstrapping technique, allow the ex ante probability of, for example, a negative GDP growth forecast for the current quarter to be estimated. The results suggest the presence of peaks of higher uncertainty related to economic recession events, with a balance of risks which became negative in the immediate aftermath of the global financial crisis.<P>Simulations stochastiques non-paramétriques pour étudier l'incertitude autour des prévisions du modèle d'indicateurs de l'OCDE<BR>L’incertitude entourant les prévisions macro-économiques ponctuelles est généralement mesurée par la performance historique du modèle de prévision, à l’aide de mesures telles que la moyenne au carré des erreurs de prévisions (EQM). Cette mesure, a cependant l’inconvénient majeur d’être constante dans le temps et donc de ne transmettre aucune information ni sur la source spécifique de l’incertitude, ni sur l’ampleur et la balance des risques liée à la conjoncture immédiate. Par ailleurs, des hypothèses paramétriques spécifiques sur la distribution de probabilité des erreurs de prévision sont nécessaires afin de dessiner des bandes de confiance autour des prévisions ponctuelles. Cet article propose une erreur quadratique moyenne simulé variant dans le temps et obtenue au moyen de simulations stochastiques nonparamétriques, combinent l’incertitude autour des paramètres du modèle et le terme d’erreurs structurelles pour construire des bandes de confiance asymétrique autour des prévisions ponctuelles. La procédure est appliquée, à titre d'exemple, aux prévisions à court terme de la croissance du PIB réel générées par le modèle d’indicateurs de l’OCDE pour l’Allemagne. Les distributions empiriques de probabilité des prévisions de croissance du PIB, obtenues par la technique de bootstrap, permettent d’estimer la probabilité ex ante d’une croissance négative du PIB pour le trimestre en cours. Les résultats suggèrent la présence de pics d’incertitude liée aux événements de la récession économique, avec une balance des risques qui est devenue négative au lendemain de la crise financière mondiale.
    Keywords: GDP, stochastic simulations, Forecasting uncertainty, empirical probability distribution, PIB, Incertitude entourant des prévisions, simulations stochastiques, distribution empirique de probabilité
    JEL: C12 C15 C53
    Date: 2012–07–27
  8. By: Wojciech Charemza; Yuriy Kharin; Vladislav Maevskiy
    Abstract: The paper aims at assessing the forecast risk and the maximum admissible forecast horizon for the non-systematic component of inflation modeled autoregressively, where a distortion is caused by a simple first-order bilinear process. The concept of the guaranteed upper risk of forecasting and the d-admissible distortion level is defined here. In order to make this concept operational we propose a method of evaluation of the p-maximum admissible forecast risk, on the basis of the maximum likelihood estimates of the bilinear coefficient. It has been found that for the majority of developed countries (in terms of average GDP per capita) the maximum admissible forecast horizon is between 5 and 12 months, while for the poorer countries it is either shorter than 5 or longer than 12. There is also a negative correlation of the maximum admissible forecast horizon with the average GDP growth.
    Keywords: Forecasting; Inflation; Bilinear Processes
    JEL: C22 C53 E31 E37
    Date: 2012–10
  9. By: Wojciech Charemza; Daniel Ladley
    Abstract: This paper considers the effectiveness of monetary policy committee voting when the inflation forecast signals, upon which decisions are based, may be subject to manipulation. Using a discrete time intertemporal model, we examine the distortions resulting from such manipulation under a three-way voting system, similar to that used by the Bank of Sweden. We find that voting itself creates persistence in inflation. Whilst altering the forecast signal, even if well intentioned, results in a diminished probability of achieving the inflation target. However, if committee members ‘learn’ in a Bayesian manner, this problem is mitigated.
    Keywords: Voting Rules; Monetary Policy; Inflation Targeting
    JEL: E47 E52 E58
    Date: 2012–10
  10. By: Àlexey Vedev (Gaidar Institute for Economic Policy)
    Abstract: Issue is devoted to an actual problem to which is still given not enough attention – a problem of long-term forecasting of national financial sector as a component of an economic complex. Problems of a forecast of development of the financial markets coordination with economic development of the country, on the one hand, and with a forecast of development of the global financial market, on the other hand are considered. Balances of the major financial instruments and institutional balances of the main sectors the economy in aggregate forming complex financial balance of the country are constructed in paper. On this basis the forecast of development of the financial markets for the period till 2020, based on assumptions of the balanced scenario is created. Besides, possibilities and transition conditions to higher growth rates of the financial market under assumption of stability are defined.
    Keywords: long-term forecasting , national financial sector, financial markets
    JEL: C01 E01 E02 E50 E61
    Date: 2012
  11. By: Domenico Giannone; Michele Lenza; Giorgio E. Primiceri
    Abstract: Vector autoregressions (VARs) are flexible time series models that can capture complex dynamic interrelationships among macroeconomic variables. However, their dense parameterization leads to unstable inference and inaccurate out-of-sample forecasts, particularly for models with many variables. A solution to this problem is to use informative priors, in order to shrink the richly parameterized unrestricted model towards a parsimonious naïve benchmark, and thus reduce estimation uncertainty. This paper studies the optimal choice of the informativeness of these priors, which we treat as additional parameters, in the spirit of hierarchical modeling. This approach is theoretically grounded, easy to implement, and greatly reduces the number and importance of subjective choices in the setting of the prior. Moreover, it performs very well both in terms of out-of-sample forecasting—as well as factor models—and accuracy in the estimation of impulse response functions.
    JEL: C11 C32 C53 E37 E47
    Date: 2012–10
  12. By: Qin, Duo (BOFIT); He, Xinhua (BOFIT)
    Abstract: Ways of extracting financial condition indices (FCI) are explored and alternative FCIs external to the Chinese economy are constructed to model their predictive content. The exploration aims at highlighting the rich and varied dynamic features of financial variables underlying FCIs and the importance of synchronising dynamic information between FCIs and the real-sector variables to be forecasted. The modelling experiment aims at improving the forecasting model upon which the FCIs are assessed. Four variables are chosen as the likely macro channel of the FCIs affecting the Chinese economy. It is found that the FCI-led models enjoy forecasting advantages over a benchmark model in three out of the four variables, although the benchmark model is not dominated by the FCI-led models when judged by in-sample encompassing tests. The evidence indicates the increasing exposure of the Chinese economy to the global financial conditions.
    Keywords: financial index; dynamic factor; VAR; error correction; encompassing
    JEL: C43 E17 F37 G17
    Date: 2012–10–18
  13. By: Mohamed Chikhi (Université de Ouargla and Université Montpellier I, Lameta); Anne Péguin-Feissolle (CNRS, Greqam); Michel Terraza (Université Montpellier I, Lameta)
    Abstract: This paper analyzes the cyclical behavior of Dow Jones by testing the existence of long memory through a new class of semiparametric ARFIMA models with HYGARCH errors (SEMIFARMA-HYGARCH); this class includes nonparametric deterministic trend, stochastic trend, short-range and long-range dependence and long memory heteroscedastic errors. We study the daily returns of the Dow Jones from 1896 to 2006. We estimate several models and we find that the coefficients of the SEMIFARMA-HYGARCH model, including long memory coefficients for the equations of the mean and the conditional variance, are highly significant. The forecasting results show that the informational shocks have permanent effects on volatility and the SEMIFARMA-HYGARCH model has better performance over some other models for long and/or short horizons. The predictions from this model are also better than the predictions of the random walk model; accordingly, the weak efficiency assumption of financial markets seems violated for Dow Jones returns studied over a long period.
    Keywords: SEMIFARMA model, HYGARCH model, nonparametric deterministic trend,kernel methodology, long memory.
    JEL: C14 C22 C58 G17
    Date: 2012–06
  14. By: Wei Sun; Robin Hanson; Kathryn Blackmond Laskey; Charles Twardy
    Abstract: A market-maker-based prediction market lets forecasters aggregate information by editing a consensus probability distribution either directly or by trading securities that pay off contingent on an event of interest. Combinatorial prediction markets allow trading on any event that can be specified as a combination of a base set of events. However, explicitly representing the full joint distribution is infeasible for markets with more than a few base events. A factored representation such as a Bayesian network (BN) can achieve tractable computation for problems with many related variables. Standard BN inference algorithms, such as the junction tree algorithm, can be used to update a representation of the entire joint distribution given a change to any local conditional probability. However, in order to let traders reuse assets from prior trades while never allowing assets to become negative, a BN based prediction market also needs to update a representation of each user's assets and find the conditional state in which a user has minimum assets. Users also find it useful to see their expected assets given an edit outcome. We show how to generalize the junction tree algorithm to perform all these computations.
    Date: 2012–10
  15. By: Alejandro Bernales; Massimo Guidolin
    Abstract: We examine whether the dynamics of the implied volatility surface of individual equity options contains exploitable predictability patterns. Predictability in implied volatilities is expected due to the learning behavior of agents in option markets. In particular, we explore the possibility that the dynamics of the implied volatility surface of individual equity options may be associated with movements in the volatility surface of S&P 500 index options. We present evidence of strong predictable features in the cross-section of equity options and of dynamic linkages between the implied volatility surfaces of equity options and S&P 500 index options. Moreover, time-variations in stock option volatility surfaces are best predicted by incorporating information from the dynamics in the implied volatility surface of S&P 500 index options. We analyze the economic value of such dynamic patterns using strategies that trade straddle and delta-hedged portfolios, and we find that before transaction costs such strategies produce abnormal risk-adjusted returns.
    Keywords: Equity options; Index options; Implied volatility surface; Predictability; Trading strategies. JEL Codes: C53, G13, G17.
    Date: 2012
  16. By: Gareth W. Peters; Alice X. D. Dong; Robert Kohn
    Abstract: Our article considers the class of recently developed stochastic models that combine claims payments and incurred losses information into a coherent reserving methodology. In particular, we develop a family of Heirarchical Bayesian Paid-Incurred-Claims models, combining the claims reserving models of Hertig et al. (1985) and Gogol et al. (1993). In the process we extend the independent log-normal model of Merz et al. (2010) by incorporating different dependence structures using a Data-Augmented mixture Copula Paid-Incurred claims model. The utility and influence of incorporating both payment and incurred losses into estimating of the full predictive distribution of the outstanding loss liabilities and the resulting reserves is demonstrated in the following cases: (i) an independent payment (P) data model; (ii) the independent Payment-Incurred Claims (PIC) data model of Merz et al. (2010); (iii) a novel dependent lag-year telescoping block diagonal Gaussian Copula PIC data model incorporating conjugacy via transformation; (iv) a novel data-augmented mixture Archimedean copula dependent PIC data model. Inference in such models is developed via a class of adaptive Markov chain Monte Carlo sampling algorithms. These incorporate a data-augmentation framework utilized to efficiently evaluate the likelihood for the copula based PIC model in the loss reserving triangles. The adaptation strategy is based on representing a positive definite covariance matrix by the exponential of a symmetric matrix as proposed by Leonard et al. (1992).
    Date: 2012–10

This nep-for issue is ©2012 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.