Quantitative Research in Financial Economics
Our group focuses on all aspects of quantitative and empirical financial modeling and draws expertise from areas as diverse as applied stochastic modeling, financial econometrics, banking, microstructure and asset pricing. The work of the group is underpinned by state-of-the-art data facilities that drive the financial risk measures contained in The Durham Finance Laboratory (coming soon).
Quantitative modeling, backed up by appropriate empirical research in relation to financial decision-making, is a topic of ongoing importance to policy-makers, financial practitioners, academics and the public at large. The mechanisms that set the interest rates on our mortgages and loans, the prices of everyday goods and services, the value of our pensions and the security of our economy relies to a greater or lesser extent on the ability to model correctly the distribution of future outcomes and hence price the degree of risk inherent within our economic system.
The Future of Quantitative Research
The various interconnected financial crises that have occurred between 2007 and 2012 have brought about a substantive change in the public understanding of the financial system. Part of our research agenda is to push forward into new areas of research relating to financial markets.
Traditional asset pricing approaches to financial markets, commonly taught in most finance courses, have focused, directly or indirectly, on the completeness argument. In a complete financial market the universe of global assets is priced relative to a finite number of states-of-the-world that have a random chance of occurring. If there are more assets than states we can use linear combinations of the states to construct prices for these assets. From this simple set-up all of our current asset pricing models have evolved, including famous models such as the capital asset pricing model and the Black and Scholes option-pricing model. The complete markets set-up is mathematically very elegant and with simple extensions makes very precise predictions on the fluctuations of financial markets.
Unfortunately, our experience during the 2007 to 2012 era has indicated that the effectiveness of these models in pricing stocks, bonds, derivatives and commodities has been very poor. Even with the introduction of more complex non-linear adjustments to the complete market approach we have not been able to capture fully the realities of the financial risks we face or the persistence of the positive and negative bubbles we have observed in various markets.
In the period leading up to 2007, financial services embraced financial engineering founded in the principles of complete markets and no-arbitrage, eschewing many of the ideas around bounded rationality and market failure that had risen to prominence in the economics literature in the prior thirty years. In the post-crisis era we are now beginning to see a re-emergence of the need to treat the agents within a financial market as strategic players with differentiated information, and this provides the backdrop for the research ideas that permeate through our group.
Part of our research agenda focuses on addressing this problem and our approach is multi-faceted. For instance, most traded underlying assets, such as stocks and commodities, have an associated derivatives market (a contract whose value is directly related to the value of the underlying asset). These markets are used to both hedge risk and leverage bets on future valuations of the underlying asset. The traditional framework provides a specific rule for the valuation of these derivatives and part of our team is involved in using the discrepancies in the predicted and actual valuations to gain insight into why our traditional asset pricing models are breaking down.
Another approach from the group focuses on the actual trading mechanism within financial markets and if the relatively new highly automated trading systems are leading to systematic inefficiencies and frictions in how we value stocks, commodities, bonds, currencies, futures, options and other financial assets. This research looks at the minutiae of trading, observing how each and every quote and who is making the quote affects the aggregate direction of asset prices. It also now appears that market participants with technological superiority in computational speed and high levels of influence appear to have systematic advantages over other participants in the market. The combination of these inefficiencies also appears to be important in explaining the inadequacies of our traditional approaches.
QRFE has an active research seminar programme. Read more
Measuring Nonlinear Granger Causality in Mean
A paper by Abderrahim Taamouti that proposes model-free measures for Granger causality in mean between economic and financial variables. The empirical results show that the variance risk premium is a very good predictor of risk premium at horizons less than six months. We also find that there is a high degree of predictability at horizon one-month which can be attributed to a nonlinear causal effect.