nep-ets New Economics Papers
on Econometric Time Series
Issue of 2007‒06‒30
eleven papers chosen by
Yong Yin
SUNY at Buffalo

  1. Non-linearities and Unit Roots in G7 Macroeconomic Variables By Yunus Aksoy; Miguel A. León-Ledesma
  2. Bayesian VARs with Large Panels By Banbura, Marta; Giannone, Domenico; Reichlin, Lucrezia
  3. What would Nelson and Plosser find had they used panel unit root tests? By Christophe Hurlin
  4. The Feldstein-Horioka Puzzle: a Panel Smooth<br />Transition Regression Approach By Julien Fouquau; Christophe Hurlin; Isabelle Rabaud
  5. Volatility forecasting for crude oil futures By Marzo, Massimiliano; Zagaglia, Paolo
  6. Conditional Leptokurtosis in Energy Prices: Multivariate Evidence from Futures Markets By Marzo, Massimiliano; Zagaglia, Paolo
  7. How useful are historical data for forecasting the long-run equity return distribution? By John M Maheu; Thomas H McCurdy
  8. Robust Bayesian Analysis of Loss Reserves Data Using the Generalized-t Distribution By Jennifer Chan; Boris Choy; Udi Makov
  9. Detecting Jumps in High-Frequency Financial Series Using Multipower Variation By Carla Ysusi
  10. Spurious Cointegration: The Engle-Granger Test in the Presence of Structural Breaks By Antonio E. Noriega; Daniel Ventosa-Santaulària
  11. Estimating Integrated Volatility Using Absolute High-Frequency Returns By Carla Ysusi

  1. By: Yunus Aksoy (School of Economics, Mathematics & Statistics, Birkbeck); Miguel A. León-Ledesma
    Abstract: We carry out a meta-analysis on the frequency of unit-roots in macroeconomic time series with a dataset covering 249 variables for the G7 countries. We use linear tests and the three popular non-linear tests (TAR, ESTAR and Markov Switching). In general, the evidence in favour of the random walk hypothesis is weaker than in previous studies. This evidence against unit roots is stronger for real and nominal asset prices. Our results show that rejection of the null of a unit root in the macro dataset is substantially higher for non-linear than linear models. Finally, the results from a Monte Carlo experiment show that rejection frequencies are very close to the nominal size of the test when the DGP is a linear unit root process. This leads us to reject the hypothesis that overfitting deterministic components explains the higher rejection frequencies of nonlinear tests.
    Keywords: overfitting, nonlinear models, unit root
    JEL: E60 C22
    Date: 2007–01
  2. By: Banbura, Marta; Giannone, Domenico; Reichlin, Lucrezia
    Abstract: This paper assesses the performance of Bayesian Vector Autoregression (BVAR) for models of different size. We consider standard specifications in the macroeconomic literature based on, respectively, three and eight variables and compare results with those obtained by larger models containing twenty or over one hundred conjunctural indicators. We first study forecasting accuracy and then perform a structural exercise focused on the effect of a monetary policy shock on the macroeconomy. Results show that BVARs estimated on the basis of hundred variables perform well in forecasting and are suitable for structural analysis.
    Keywords: Bayesian VAR; forecasting; large cross-sections; monetary VAR
    JEL: C11 C13 C33 C53
    Date: 2007–06
  3. By: Christophe Hurlin (LEO - Laboratoire d'économie d'Orleans - [CNRS : UMR6221] - [Université d'Orléans])
    Abstract: In this study, we systemically apply nine recent panel unit root tests to the same fourteen macroeconomic and financial series as those considered in the seminal paper by Nelson and Plosser (1982). The data cover OECD countries from 1950 to 2003. Our results clearly point out the difficulty that applied econometricians would face when they want to get a simple and clear-cut diagnosis with panel unit root tests. We confirm the fact that panel methods must be very carefully used for testing unit roots in macroeconomic or financial panels. More precisely, we find mitigated results under the cross-sectional independence assumption, since the unit root hypothesis is rejected for many macroeconomic variables. When international cross-correlations are taken into account, conclusions depend on the specification of these cross-sectional dependencies. Two groups of tests can be distinguished. The first group tests are based on a dynamic factor structure or an error component model. In this case, the non stationarity of common factors (international business cycles or growth trends) is not rejected, but the results are less clear with respect to idiosyncratic components. The second group tests are based on more general specifications. Their results are globally more favourable to the unit root assumption.
    Keywords: Panel Unit Root Tests
    Date: 2007–06–22
  4. By: Julien Fouquau (LEO - Laboratoire d'économie d'Orleans - [CNRS : UMR6221] - [Université d'Orléans]); Christophe Hurlin (LEO - Laboratoire d'économie d'Orleans - [CNRS : UMR6221] - [Université d'Orléans]); Isabelle Rabaud (LEO - Laboratoire d'économie d'Orleans - [CNRS : UMR6221] - [Université d'Orléans])
    Abstract: This paper proposes an original framework to determine the relative influence of five<br />factors on the Feldstein and Horioka result of OECD countries with a strong saving-<br />investment association. Based on panel threshold regression models, we establish<br />country-specific and time-specific saving retention coefficients for 24 OECD coun-<br />tries over the period 1960-2000. These coefficients are assumed to change smoothly,<br />as a function of five threshold variables, considered as the most important in the<br />literature devoted to the Feldstein and Horioka puzzle. The results show that; de-<br />gree of openness, country size and current account to GDP ratios have the greatest<br />influence on the investment-saving relationship.
    Keywords: Feldstein Horioka puzzle, Panel Smooth Threshold Regression models,<br />saving-investment association, capital mobility .
    Date: 2007–06–22
  5. By: Marzo, Massimiliano (Department of Economics, Universit`a di Bologna); Zagaglia, Paolo (Dept. of Economics, Stockholm University)
    Abstract: This paper studies the forecasting properties of linear GARCH models for closing-day futures prices on crude oil, first position, traded in the New York Mercantile Exchange from January 1995 to November 2005. In order to account for fat tails in the empirical distribution of the series, we compare models based on the normal, Student’s t and Generalized Exponential distribution. We focus on out-of-sample predictability by ranking the models according to a large array of statistical loss functions. The results from the tests for predictive ability show that the GARCH-G model fares best for short horizons from one to three days ahead. For horizons from one week ahead, no superior model can be identified. We also consider out-of-sample loss functions based on Value-at-Risk that mimic portfolio managers and regulators’ preferences. EGARCH models display the best performance in this case.
    Keywords: GARCH models; kurtosis; oil prices; forecasting
    JEL: C22 G19
    Date: 2007–06–21
  6. By: Marzo, Massimiliano (Department of Economics, Universit`a di Bologna); Zagaglia, Paolo (Dept. of Economics, Stockholm University)
    Abstract: We study the joint movements of the returns on futures for crude oil, heating oil and natural gas at a daily frequency. We model the leptokurtic behavior through the multivariate GARCH with dynamic conditional correlations and elliptical distributions introduced by Pelagatti and Rondena (2004). Futures prices of crude and heating oil co-vary strongly. The correlation between the futures prices of natural gas and crude oil has been rising over the last 5 years. However, this correlation has been low on average over two thirds of the sample, indicating that futures markets have no established tradition of pricing natural gas as a function of developments on oil markets.
    Keywords: Multivariate GARCH; Kurtosis; Energy Prices; Futures Markets
    JEL: C22 G19
    Date: 2007–06–27
  7. By: John M Maheu; Thomas H McCurdy
    Abstract: We provide an approach to forecasting the long-run (unconditional) distribution of equity returns making optimal use of historical data in the presence of structural breaks. Our focus is on learning about breaks in real time and assessing their impact on out-of-sample density forecasts. Forecasts use a probability-weighted average of submodels, each of which is estimated over a different history of data. The paper illustrates the importance of uncertainty about structural breaks and the value of modeling higher-order moments of excess returns when forecasting the return distribution and its moments. The shape of the long-run distribution and the dynamics of the higher-order moments are quite different from those generated by forecasts which cannot capture structural breaks. The empirical results strongly reject ignoring structural change in favor of our forecasts which weight historical data to accommodate uncertainty about structural breaks. We also strongly reject the common practice of using a fixed-length moving window. These differences in long-run forecasts have implications for many financial decisions, particularly for risk management and long-run investment decisions.
    Keywords: density forecasts, structural change, model risk, parameter uncertainty, Bayesian learning, market returns
    JEL: C51 C53 C11
    Date: 2007–06–28
  8. By: Jennifer Chan (University of Sydney); Boris Choy (Department of Mathematical Sciences, University of Technology, Sydney); Udi Makov (University of Haifa)
    Abstract: This paper presents a Bayesian approach using Markov chain Monte Carlo methods and the generalized-t (GT) distribution to predict loss reserves for the insurance companies. Existing models and methods cannot cope with irregular and extreme claims and hence do not offer an accurate prediction of loss reserves. To develop a more robust model for irregular claims, this paper extends the conventional normal error distribution to the GT distribution which nests several heavytailed distributions including the Student-t and exponential power distributions. It is shown that the GT distribution can be expressed as a scale mixture of uniforms (SMU) distribution which facilitates model implementation and detection of outliers by using mixing parameters. Different models for the mean function, including the log-ANOVA, log-ANCOVA, state space and threshold models, are adopted to analyze real loss reserves data. Finally, the best model is selected according to the deviance information criterion (DIC).
    Keywords: Bayesian approach; state space model; threshold model; scale mixtures of uniform distribution; device information criterion
    Date: 2007–05–01
  9. By: Carla Ysusi
    Abstract: When the log-price process incorporates a jump component, realised variance will no longer estimate the integrated variance since its probability limit will be determined by the continuous and jump components. Instead realised bipower variation, tripower variation and quadpower variation are consistent estimators of integrated variance even in the presence of jumps. In this paper we derive the limit distributions of realised tripower and quadpower variation, allowing us to compare these three estimators of integrated variance. Using the limit theories for the differences of the errors, tests for jumps are proposed for each estimator. Using simulated data, the performance of each of these tests is compared. The tests are also applied to empirical data but results need to be taken carefully as market microstructure effects may contaminate real data.
    Keywords: Quadratic variation, Multipower variation, Stochastic volatility models, Jump process, Semimartingale, High-frequency data
    JEL: C12 C51 G19
    Date: 2006–09
  10. By: Antonio E. Noriega; Daniel Ventosa-Santaulària
    Abstract: This paper analyses the asymptotic behavior of the Engle-Granger t-test for cointegration when the data include structural breaks, instead of being pure I(1) processes. We find that the test does not possess a limiting distribution, but diverges as the sample size tends to infinity. Calculations involving the asymptotic expression of the t-test , as well as Monte Carlo simulations, reveal that the test can diverge in either direction, making it unreliable as a test for cointegration, when there are neglected breaks in the trend function of the data. Using real data on car sales and murders in the US, we present an empirical illustration of the theoretical results.
    Keywords: Spurious cointegration, structural breaks, integrated processes
    JEL: C12 C13 C22
    Date: 2006–12
  11. By: Carla Ysusi
    Abstract: When high-frequency data is available, in the context of a stochastic volatility model, realised absolute variation can estimate integrated spot volatility. A central limit theory enables us to do filtering and smoothing using model-based and model-free approaches in order to improve the precision of these estimators. Although the absolute values are empirically attractive as they are less sensitive to possible large movements in high-frequency data, realised absolute variation does not estimate integrated variance. Some problems arise when using a finite number of intra-day observations, as explained here.
    Keywords: Quadratic variation, Absolute variation, Stochastic volatility models, Semimartingale, High-frequency data
    JEL: C13 C51 G19
    Date: 2006–12

This nep-ets issue is ©2007 by Yong Yin. 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.