It is widely observed by markets participants that during periods of high market volatility, correlations between asset prices can differ substantially from those seen in period of low market volatility. For example, correlations between US Equities represented by S&P 500 and energy commodities as crude oil, gasoline, natural gas; were substantially lower prior the bull market started in 2003. Such differences in correlations among asset classes could highly affect portfolio returns and reduce the benefits of diversification, the time mostly desired by investors. The change in correlations had been attributed either to structural breaks in the underlying distribution of returns or to "contagion" across markets that occurs only during periods of market turbulence. This will be examined in order the above mentioned practitioner's observations hold, and whether are statistically significant.
Motivation The reasons for this research are mainly based on the empirical observations of market participants. With this as a starting point I would like to examine whether these observations stand and whether are statistically significant. Moreover, another reason for this research is that I am interested in asset management and I would like to deepen my knowledge regarding issues that are being faced by professional asset managers; and the implications in their investment decision making procedure. Furthermore, during the recent financial crisis, starting on January 2008 there was noticed by investors that the so called decoupling between European and Asian emerging stock markets is not that robust as they were thinking. The rational that the growth rate of European, Asian and South American emerging markets have broadened and deepened to the point that they no longer depend on the United States economy for growth, leaving them insulated from a severe slowdown there. Faith in the concept has generated strong outperformance for stocks outside the United States. However, in January 2008, as fears of recession mounted in the United States, stocks declined heavily. Contrary to what the supporters of the decoupling theory would have expected, the losses were greater outside the United States, with the worst experienced in emerging markets.
Literature Review Loretan and English (2000), extend previous empirical studies, allowing for robustly estimated, time-varying coeÂ¬Æ’cients. However, there are no existing studies of aggregated credit spreads, stocks and stock market volatility in which the conditional covariance structure is considered. Therefore their Â¬Ândings oÂ¬â‚¬er some of the Â¬Ârst insights regarding the variance and correlation structure of this data set. They found evidence of strongly varying conditional variances and correlations, with dependencies increasing after the outbreak of the Â¬Ânancial crisis. This knowledge opens the door to better decision tools in various areas, such as asset pricing, portfolio selection, and risk management. Their conclusion is in line with the empirical suggestions and my expected results. Another study of Chong and Miffre (2008); this studied the correlations among 25 different commodities and 7 global stock indices and 6 bond indices, found that the correlations between S&P500 and 19 of the 25 examined commodities decreased during periods of increased volatility. This is a very interesting finding as it suggests that assets managers and investors can really benefit from diversification and its benefits as during turbulent times commodities and stock indices tend to correlate less. This does not apply to all the commodities examined. For example for crude oil, live cattle, unleaded gas and coffee the correlation with S&P amplified in periods of turmoil. Also even for the rest commodities that the correlations decrease, this does not apply to the same degree. More specifically for the precious metals as gold, silver and platinum the correlations with S&P500 drop much more than the rest commodities in turbulent times.
Description of Methodology and Data The data will be collected from WRDS Database and will comprise S&P500 Index, Russell 2000, Eurostoxx 50, selected US Treasury Notes and Bonds, commodity indexes like CRB Commodity Index, as well as specific commodity future contracts. Moreover I plan to include emerging markets indices which are indicative of high growth emerging markets like the CSI 300 index for Chinese companies, the Bovespa index for Brazilian companies and the NIFTY index representing the biggest Indian companies. Finally I would also include in my study certain currency pairs, as the EUR/USD, the CAD/USD and the USD/JPY. The reason for this decision, to include specific currency pairs and consider them as an asset class; is that the past years more and more institutional investors and many sovereign funds invested in currencies other than their country of origin in order to benefit from fluctuations in exchange rates or for diversification. Second reason for the selection of currency pairs as a separate asset class is the very high volatility experienced the past decade even between major currencies like EUR, USD, GBP, CHF, and JPY. The frequency of the data will be weekly for the time period starting on January 1994 till May 2011. In order to avoid the weekend effect, thin trading effect and maturity of future contract effect, Wednesday's settlement prices will be considered as the closing weekly prices.
Expected Empirical Results After the conclusion of the study, I expect to obtain quantitative results about in what extent change the correlations of different asset classes, which have mentioned above, with each other and whether these respective changes are statistically significant from low to high volatile periods. Also I expect to examine whether certain asset managers practices of the market really hold, through statistical significance tests. These are widely practices of asset market participants. One usually stated, is that when stocks decline then US Government Bonds and gold prices increase, as investors look for a safe heaven. In addition I expect to clarify whether these market tendencies intensify, decline or remain stable from crisis to crisis as the time passes. I expect to find that commodities prices are not that strongly correlated with major market indices and tend to retain their independence even in crisis periods. But this holds in different degree and not for all the commodities. Furthermore, I expect correlations between commodities and stock indices to change significant because of asset management trends. I expect to find less correlation during previous crises compared to the recent financial crisis where many commodities were perceived as a separate immunized from market fluctuations, asset class, resulting a bubble in certain commodities as well. Finally I expect that certain currency pairs are more correlated with the stock indices. For instance USD is considered to be a safe heaven and during periods of market turmoil it is less correlated with the stock markets or even negatively correlated as the stock markets decline its values compared to other currencies increases as investors desire to hold more cash on hand and especially US dollars. In addition to that I expect the CAD / USD to be more correlated with the stock markets the period from 2003 till 2008 where the advance in stock markets was partly driven from good performing mining companies, which were performing well partly because of high commodities prices.