160712.pdf

Abstract

A flash crash is an extremely rapid movement in security prices that happens over a very short period of time. With the advent of trading algorithms, trades can be executed within milliseconds in the stock market. More often than not, the inter connectivity of these trading algorithms are blamed to be a part of the reason behind flash crashes. If a flash crash lasts long enough, then it can signal loss of confidence through the market. This could mean anything from a loss for investors to a trigger for recession (if it happens at the wrong time of a business cycle). Since the financial crisis of 2007-09, modelling the spread of systemic risk within a network of financial institutions has been a topic of interest. SIR models are widely used to model the spread of an infectious disease in epidemiology. These models help determine the number of susceptible, infected and recovered individuals in a population that further helps in evaluating the number of people that need to be vaccinated to prevent an outbreak among other essential uses. Parallels have been drawn between the spread of systemic risk in banking ecosystems and the spread of infectious diseases with an aim to minimize the risk. Such spread modelling approaches have also been applied to a network of financial institutions in order to measure the potential for contagion and using contagion as a means to model and find out indicators for financial crises. This thesis sets out to analyse and propose a solution for the flash crashes which, as described above, pose a significant threat to markets. As a first step, the qualitative and quantitative indicators of flash crashes will be listed. Followed by that, spread modelling approaches and prior literature on the use of similar approaches to model systemic risk within financial markets will form the fundamental lens to look at the problem.