Risk Connectedness: Evolution, Intervention, and Application

Abstract

In recent years, global uncertainty has surged due to the multi-effects of COVID-19, geopolitical conflicts, energy price fluctuation, downward pressure on the economy, inflation, policy uncertainty, and natural disasters. Meanwhile, the fast growth of emerging regions and new financial instruments has paved the way for potential investors to enjoy a large set of diversification opportunities and higher returns. In this context, the research of risk spillover in asset markets should aim at not only preventing financial crisis but also stimulating market vitality, fostering green investment, and nurturing emerging markets. My research studies the investment concentration of assets and procyclical effects by assessing risk connectedness in asset markets. A common thread of my PhD thesis has been applying the connectedness measure of Diebold & Yilmaz (2014) to international financial markets, following a ¿what-why-how¿ approach. My work aims at interpreting observed evolutions of connectedness (what), studying driving forces and intervention policies of connectedness (why), and fostering better market participation (how). My research covers a range of issues in financial risk management, including crises in developed and emerging markets, prudential policies, green investments, new econometric tools, and derivative innovation. The PhD thesis is divided into three parts. The first part measures risk connectedness among assets in the uncertain world. This measurement covers as many countries and asset classes as possible, including (1) during recent three decades with many historical events such as crisis & war, I study the four asset classes of stock, currency, commodity, and bond in U.S. market; (2) I study the relationship between green and black bonds in six countries and territories in Asia before and after COVID-19; (3) I study 68 companies from four energy-related sectors from 2011 to 2019. These results show existed pattern of spillover risk in different regions and asset classes. For example, results reveal that stocks and commodities often have high spillover risk to other assets during short-term fluctuations, while the shocks of currencies have more significant impacts on others in long term. The second part of this thesis addresses the relationship between connectedness and factors such as prudential policy, national macroeconomy, and credit rating changes, etc. I focus on the risk connectedness in ASEAN-4 sovereign market, because of the increased issuance amount of treasury and the unignorable global uncertainty¿s impacts on the ASEAN-4 market. To the best of my knowledge, the intervention of network spillover risk in sovereign bond market vulnerability has not been explored in previous literature. I find that the key drivers of connectedness include VIX, Trade Balance, Government Bond Issuance, geopolitical risk, and COVID-19. One of the interesting findings is that the prudential policy has both direct and indirect effects on the spillover risk from other countries to ASEAN-4 treasuries. The direct effect is that markets with tighter prudential policies face significantly smaller spillovers from the treasury yield shocks of other regional and global countries. While the economic growth and credit rating upgrade significantly caused by the prudential policy will aggravate the vulnerability of sovereign bonds, giving rise to a weakened effect of the prudential policy. The sum of indirect and direct effects indicates that prudential policies reduce sovereign spillover risks in the long term. These findings suggest prudential policies have dual efficiency in sovereign risk regulation and treasury internationalization. Additionally, I find that ASEAN bonds injected a positive influx into the three developed markets during global economic turmoil. The third part of this thesis intends to provide regulators and investors with two new practical innovations in risk management. The first practical innovation is a new econometric model. I utilize the intraday financial data and construct a parsimonious model that introduces a network factor to Realized GARCH of Hansen et al. (2012). This model, called the RVNET-GARCH model, satisfies the practical demand from risk regulators and portfolio managers for synthesizing the consequences of the spillover risks from multiple sources to one target asset¿s volatility. Within the RVNET-GARCH model, the news impact curve is extended to the multivariate setting, and an asymmetric measure of the news-to-volatility interconnectedness is also proposed in a multivariate network. By utilizing the stock indices of 22 countries and territories, I evaluate the RVNET-GARCH¿s out-of-sample volatility forecasting performances against Realized GARCH model of Hansen et al. (2012). Results show that the new model has higher accuracy and lower predicting error for all samples. The second practical innovation is about financial instruments. Fostering the widespread adoption of environmentally sustainable renewable energy resources (Clean Energy) is one of the most important challenges the world is facing today. The existence of financing vehicles such as derivative instruments in clean energy markets analogous to those commonly found in the traditional energy markets would have the capacity to transfer risk and facilitate greater liquidity. I study green bond volatility swap (GBVS) to mitigate volatility uncertainty in going green. In GBVS pricing, I utilize the RVNET-GARCH model which has been introduced in last section. The RVNET-GARCH has natural superiority because it utilizes realized variance in fitting and forecasting, while the payoff of GBVS is measured by the gap between strike volatility and realized volatility as well. By testing 32 green bonds in China from Jan 2019 to May 2022, the new pricing model robustly outperforms the classical benchmark model of Javaheri et al. (2004) in payoff errors. These results will be of interest to market actors and decision-makers in the asset markets. Indeed, from the 1990s to the 21st century, the intensified connectedness in the asset market caused a weak self-healing capability of financial markets. In this sense, the novel developments of derivative and econometric tools contain key enhancements to crisis prevention as well as long-term market vitality.