Special issue of the International Journal of Finance and Economics innovations in finance, economics, risk management, and policy

• Author: Georgios Sermpinis,Athanasios A. Pantelous,Keith Cuthbertson,Ioannis Kyriakou
• Published date: 01 October 2019
• DOI: http://doi.org/10.1002/ijfe.1738
• Date: 01 October 2019

DOI: 10.1002/ijfe.1738

EDITORIAL

Special issue of the International Journal of Finance and

Economics innovations in finance, economics, risk

management, and policy

This special issue comprises a selection of papers cover-

ing a broad rangeof finance topics from the 4th Symposium

on Quantitative Finance and Risk Analysis (QFRA) held

in June 2018 at the island of Mykonos in Greece.

Two papers in this special edition look at the infor-

mation content of options' prices. Measuring Value at

Risk (VaR) and Conditional Value at Risk (CVaR) using

statisticalmethodssuchasARCHandGARCHiswell

established. A key limitation of such methods is the

backward-looking nature of this “statistical approach”

and the fact that the most recent values of the vari-

ables are not given much weight in forecasting future

values. The paper by Barone-Adesi, Legnazzi, and Sala

(Option-implied risk measures: An empirical examination

on the S&P500 index) uses data from index options to

extract forward-looking measures of volatility for use in

VaR calculations. The authors' back-test their VaR fore-

casts using statistical approaches and the “option-implied”

approach (on the S&P500 index) but do not find huge

differences between them. However, the option-implied

approach is simple to implement and is found to be at

least as accurate as the more complex statistical approach.

Therefore, the implied-options approach might be a use-

ful complement to existing statistical methods for internal

and external validation of VaR forecasts.

The information in options data is also used in the

paper by Voukelatos and Verousis (Option-implied infor-

mation and stock herding) to examine herding behaviour

in equity markets. The novel approach in this paper is to

link herding in the US equity market (daily data, January

1996–December 2015) with information from the options

market, such as (index) implied volatility, implied volatil-

ity skew, and implied volatility spreads. Using data on a

cross-section of US stocks, herding is found to be stronger

during periods of market stress.

The next set of four papers cover the key areas

of portfolio allocation, IPOs, mergers and acquisitions,

and the determinants of changes in bond prices. The

mean-variance approach to portfolio allocation in its basic

form has well-known practical drawbacks, namely, the

sensitivity of optimal weights to poor forecasts of the

inputs such as mean returns and second moments. The

paper by Zhao, Stasinakis, Sermpinis, and Da Silva Fernan-

des (Revisiting Fama-French factors' predictability with

Bayesian modelling and copula-based portfolio optimiza-

tion) attempts to remedy such defects. They take the

five Fama-French US (style) factors as their basic return

inputs and undertake a three-stage optimisation proce-

dure. First, they use several linear and non-linear models

to forecast expected returns, which are then combined

using a Bayesian shrinkage approach and back-tested

to ascertain their forecasting accuracy. Second, they use

an asymmetric copula model to forecast the second

moments of the factors. Finally, the optimal asset weights

are determined using both the mean-variance approach

and the mean-variance CVaR criteria. The performance

metrics used to assess success include the Sharpe and

Sortino ratios and maximum drawdown. They find, in

the out-of-sample period 2000–2017 using monthly data,

that their three-stage optimisation approach performs

well in comparison to other previously used methods,

such as forecasting returns using a random walk or the

1/N-portfolio approach.

The paper by Sermpinis, Tsoukas, and Zhang (What

influences a bank's decision to go public?) looks at the

determinants of IPOs using US data over the period 1996

to 2016. The candidate variables to help predict IPOs con-

sist of bank-specific variables, such as size, profitability,

capital-assets ratio, and leverage, as well as macroeco-

nomic variables, such as interest rates and GDP growth.

Several candidate models, including the Cox proportional

hazard, discrete hazard, and logistic models are analyzed.

But the methodological innovation is the application of a

“least absolute shrinkage and selection operator” (LASSO)

regression approach, allowing general-to-specific mod-

elling of the many candidate independent variables, by

forcing some coefficients of these variables to zero and

shrinking others. The authors find that this consider-

ably helps improve the ability of the model to forecast

future IPOs.

The determinants and relative success, or otherwise,

of Mergers & Acquisitions (M&A) has been widely stud-

ied using the Cumulative Abnormal Return (CAR) as a

measure of performance. The paper by Huang, Zhang,

Int J Fin Econ. 2019;24:1407–1408. wileyonlinelibrary.com/journal/ijfe © 2019 John Wiley & Sons, Ltd. 1407