"Volatility Forecasting and asset allocation in portfolio management"
Lead Research Organisation:
University of Kent
Department Name: Kent Business School
Abstract
My proposal area of study is Volatility Forecasting considering all the alternative
definitions of volatility. Each supports a different objective and all are equally
significant. Volatility is the conditional standard deviation of the asset returns and
suggests the random factor of the time series.
Risk is fundamental on financial decision making process. Thus volatility forecasting
area has attracted many researchers. Many famous financial models require
calculating the volatility. Illustrations are the Value-at Risk and Expected Shortfall
models, asset allocation in portfolio management using the Markowitz method and
asset and derivative pricing (Black-Scholes formula). However, volatility's utilization
spreads to other areas of social sciences as well.
The main characteristic of volatility (and covariance) is that is unobservable even expost.
Thus, producing accurate out-of-sample forecasts is vital.
The first part of the study will focus on model estimation and volatility forecasting.
The models are mainly classified into univariate and multivariate ones. Following
forecasting volatility, evaluation of those forecasts will be performed to test whether
the forecasts are adequate enough. The second part will focus on asset allocation
issues within a portfolio.
definitions of volatility. Each supports a different objective and all are equally
significant. Volatility is the conditional standard deviation of the asset returns and
suggests the random factor of the time series.
Risk is fundamental on financial decision making process. Thus volatility forecasting
area has attracted many researchers. Many famous financial models require
calculating the volatility. Illustrations are the Value-at Risk and Expected Shortfall
models, asset allocation in portfolio management using the Markowitz method and
asset and derivative pricing (Black-Scholes formula). However, volatility's utilization
spreads to other areas of social sciences as well.
The main characteristic of volatility (and covariance) is that is unobservable even expost.
Thus, producing accurate out-of-sample forecasts is vital.
The first part of the study will focus on model estimation and volatility forecasting.
The models are mainly classified into univariate and multivariate ones. Following
forecasting volatility, evaluation of those forecasts will be performed to test whether
the forecasts are adequate enough. The second part will focus on asset allocation
issues within a portfolio.
Organisations
Studentship Projects
| Project Reference | Relationship | Related To | Start | End | Student Name |
|---|---|---|---|---|---|
| ES/P00072X/1 | 01/10/2017 | 30/09/2027 | |||
| 1938235 | Studentship | ES/P00072X/1 | 01/10/2017 | 21/04/2021 | Eirini Bersimi |
| Description | We have been working on the area of volatility forecasting and evaluation. Based on the recent findings that forecast combinations outperform single models, a hierarchical procedure for combining volatility forecast was applied. We proved that this approach works when forecasting stock return volatility as well. Then the forecasts and the combinations were evaluated for their ability to predict the direction of change of the variable of interest. The main results show that forecast combinations considered are highly competitive against the best performer GARCH model in terms of directional accuracy and directional value. These results seem to suggest that when there is uncertainty about what model to be chosen, the methods of combination may be preferable in case of directional forecast equivalence among models and forecast combinations. Then we focus on an application of multivariate volatility models. This study aims at constructing, risk-based, Risk Parity (RP) portfolios of hedge funds which are competing against the Hedge Fund Research (HFR) RP Indices. Our results suggest our constructed portfolio have a better performance than the benchmark Indices. Finally, given the lack of agreement in the literature, we examine under what circumstances, option-Implied Volatility models provide more accurate density forecasts compared to historical volatility. What is more, since densities implied by option prices produce unrealistic risk-neutral densities (RND) calibrated real-world densities (RWD) are also calculated. We found that historical densities outperform RND and RWD forecasts. |
| Exploitation Route | It contributes to the literature by answering questions that have not be addressed yet. The findings could be also be used as evidence for future research. |
| Sectors | Financial Services, and Management Consultancy |
| Description | Overseas Institutional Fund |
| Amount | £2,050 (GBP) |
| Organisation | Economic and Social Research Council |
| Sector | Public |
| Country | United Kingdom |
| Start | 10/2019 |
| End | 12/2019 |