Search

The availability of numerous modeling approaches for volatility forecasting leads to model uncertainty for both researchers and practitioners. A large number of studies provide evidence in favor of combination methods for forecasting a variety of financial variables, but most of them are implemented on returns forecasting and evaluate their performance based solely on statistical evaluation criteria. In this paper, we combine various volatility forecasts based on different combination schemes and evaluate their performance in forecasting the volatility of the S&P 500 index. We use an exhaustive variety of combination methods to forecast volatility, ranging from simple techniques to time‐varying techniques based on the past performance of the single models and regression techniques. We then evaluate the forecasting performance of single and combination volatility forecasts based on both statistical and economic loss functions. The empirical analysis in this paper yields an important conclusion. Although combination forecasts based on more complex methods perform better than the simple combinations and single models, there is no dominant combination technique that outperforms the rest in both statistical and economic terms.

This paper examines the dynamic relationship between stock and bond returns in eleven Eurozone countries during the last seventeen years. The literature so far reports heterogeneous results with respect to the important determinants of the stock-bond relationship. To deal with model uncertainty we employ a Bayesian model averaging technique and examine various macroeconomic and financial variables which are likely to influence stock-bond comovement. Bond and stock market uncertainty, interest rate, inflation and state of the economy are important determinants of cross-asset correlations. Divergence in the dynamic patterns of stock-bond comovement as well as on the effect of economic variables on this comovement is reported during crisis periods and between different European regions. Our results are of high relevance for investment strategies as well as for policy decisions in the European context.

This study explores the dynamic properties and predictability of the Realized Minimum Variance Hedge Ratio (RMVHR), constructed from five-minute spot and future returns of two stock indices and two exchange rates. A number of econometric models are employed to forecast directly the RMVHR and the out-of-sample performance is evaluated. Results from statistical measures suggest that the evolution of the realized hedge ratio series is predictable. In terms of risk reduction, we conclude that realized hedge ratio forecasts dominate conventional methods that use daily data while the benefit is pronounced when economic gains are considered. The superior performance of RMVHR methods holds across different asset classes but is more conspicuous in the case of stock indices. Finally, this study assesses the effect of sampling frequency and transaction costs.

This paper investigates the existence of predictable patterns in the evolution of the implied correlation series. To this end, alternative time-series specifications are employed to model the correlation dynamics, and the statistical and economic significance of out-of sample forecasts is assessed. The statistical measures provide strong evidence in favor of predictable patterns in the SP however, these profits disappear when transaction costs are incorporated. We conclude that the efficient market hypothesis cannot be rejected.

Trading activity in G7 stock markets reflects not only the macroeconomic and financial impact of these G7 economies in international economic growth, but also their financial interdependence. While this nexus of major stock markets has been explored in terms of volatility and return spillovers, there has been no combined analysis of return, volatility and illiquidity spillovers. We study illiquidity spillovers because they are transmissions of trading activity and, thereof, transmissions of information and market sentiment. We find that the dynamics of international stock markets are characterized by persistent illiquidity and also that illiquidity shocks are significantly correlated across markets. Furthermore, we discover Granger causal associations between risk, return and illiquidity across G7 stock market and also within each stock market. Our findings bear significance for the regulation of international financial markets and also for international portfolio diversification.

We propose to adopt high-frequency DCC-MIDAS models to estimate high- and low-frequency correlations in the 10-year government bond spreads for Belgium, France, Italy, the Netherlands, and Spain relative to Germany, from June 1, 2007 to May 31, 2012. The high-frequency component, reflecting financial market conditions, is evaluated at 15-minute frequency, while the low-frequency component, fixed through a month, depends on country-specific macroeconomic conditions. We find strong links between spreads volatility and worsening macroeconomic fundamentals; in presence of similar macroeconomic fundamentals relative spreads move together; the increasing correlation in spreads during the burst of the sovereign debt crisis cannot be entirely ascribed to macroeconomic factors but rather to changes in market liquidity.

Reliable correlation forecasts are of paramount importance in modern risk management systems. A plethora of correlation forecasting models have been proposed in the open literature, yet their impact on the accuracy of value-at-risk calculations has not been explicitly investigated. In this paper, traditional and modern correlation forecasting techniques are compared using standard statistical and risk management loss functions. Three portfolios consisting of stocks, bonds and currencies are considered. We find that GARCH models can better account for the correlation's dynamic structure in the stock and bond portfolios. On the other hand, simpler specifications such as the historical mean model or simple moving average models are better suited for the currency portfolio. Copyright © 2007 John Wiley & Sons, Ltd.

Value at risk (VaR) is an estimate of the worst loss over a target horizon with a given level of confidence. Despite its shortcomings, the Basel Committee on Banking Supervision has chosen it as the standard method to measure the market risk of a portfolio of financial assets. This measure of risk is widely used by practitioners and regulators because of its conceptual simplicity and flexibility. This article examines the systematic relationship between correlation misestimation and the corresponding value-at-risk miscalculation. To this end, first a semi-parametric approach, and then a parametric approach is developed. Both approaches are based on a simulation setup. Various linear and non-linear portfolios are considered, as well as variance-covariance and Monte-Carlo simulation methods are employed. Results show that the VaR error increases significantly as the correlation error increases, particularly in the case of well-diversified linear portfolios. In the case of option portfolios, this effect is more pronounced for short-maturity, in-the-money options. The use of MC simulation to calculate VaR magnifies the correlation bias effect. These results have important implications for measuring market risk accurately.

One of the inputs required by investors, seeking to hold efficient portfolios, is the correlation between the securities to be included in the portfolio. Correlation estimates are required in most applications in finance including asset pricing models, capital allocation, risk management and option pricing and hedging. This paper presents an overview of my PhD thesis that focused on developing and investigating innovative methods for estimating and forecasting correlation between financial asset returns. Firstly, the importance of correlation estimation in financial applications is examined focusing on a risk management application. Secondly, a traditional financial economics problem regarding the international financial market linkages is addressed based on a dynamic correlation structure. An application to the European bond markets volatility transmission mechanism is provided. Thirdly, an innovative methodology for forecasting correlation using the market prices of derivative instruments is developed.

Most approaches in forecasting future correlation depend on the use of historical information as their basic information set. Recently, there have been some attempts to use the notion of “implied” correlation as a more accurate measure of future correlation. This study proposes an innovative methodology for backing-out implied correlation measures from index options. This new measure called implied correlation index reflects the market view of the future level of the diversification in the market portfolio represented by the index. The methodology is applied to the Dow Jones Industrial Average index, and the statistical properties and the dynamics of the proposed implied correlation measure are examined. The evidence of this study indicates that the implied correlation index fluctuates substantially over time and displays strong dynamic dependence. Moreover, there is a systematic tendency for the implied correlation index to increase when the market index returns decrease and/or the market volatility increases, indicating limited diversification when it is needed most. Finally, the forecast performance of the implied correlation index is assessed. Although the implied correlation index is a biased forecast of realized correlation, it has a high explanatory power, and it is orthogonal to the information set compared to a historical forecast. © 2005 Wiley Periodicals, Inc. Jrl Fut Mark 25:171–197, 2005