Work Package N.4 is the most applied of the project, and is mainly dedicated to the rising role and importance of High Frequency Traders (HFT) in financial markets. Biais and Foucault (2014), and the SEC (2014) survey the young and rapidly growing high-frequency trading literature. WP4 will thus borrow from the new theoretical findings of WP1, WP2 and WP3 to study high-frequency trading from a new perspective. The work package is composed by three lines of research.
 
1) The fisrt line of research of WP4 focuses on the empirical analysis of flash crashes and their relationship with algorithmic trading and HFTs. In particular, we will investigate the role of HFT during flash crashes by distinguishing between different types of HFT. In fact, Investment Banks that adopt algorithmic trading might have a different role with respect to the “pure” HFT on affecting the price and the market liquidity during flash crashed because of their different capital budget and inventory capacity. As well, both Investment Banks HFT or “pure” HFT might behave differently if they trade in their own account or if they serve as designated market makers.
This trader classification is available foracademics thanks to the BEDOFIH database for the NYSE Euronext Paris exchange made available by Eurofidai. The literature on flash crashes and the role ofHFT is vast but largely focuses either on the single event of 2010 or on a very limited number or events (as in Kirilenko et al., 2017, or Hautsch, Noé andZhang, 2017, respectively). Instead, by using the approach proposed by Christensen, Oomen and Renò (2017), we plan to identify several flash crashes in the data and then investigate the persistency in the heterogeneous behaviours of the different HFTs.
2) The second line of research of WP4 focuses and the theoretical and empirical investigation of “market resilience” to flash crashes in presence of HFT. A significant aspect of analysing financial stability relates to the nonlinear dynamics of the financial system. Nonlinear systems can exhibit very large changes in behaviour as a result of very small alterations of key parameters or variables. In particular Lo (2017) in his book “Adaptive Markets Financial Evolution at the Speed of Thought”, by reconciling the fragility of financial markets with market efficiency and rational expectation, provide insights on when nonlinearities may play a crucial role in financial markets (see also Brennan, Lo 2014, and Dindo and Massari, 2017, for preliminary formal applications of these ideas). Nonlinearities may be crucial in presence of HFT which link different stock exchanges. A liquidity shock on one venue, that might have gone unnoticed if there were one large centralized exchange, can instead affect prices when the venue is a separate exchange. In normal times, the aberrant price would quickly disappear as cross-trading-venue HFT would profit from buying low and sell high, and thus provide the needed liquidity. However, in stressed markets their capital may be limited, or the traders themselves may start to doubt the prices (as happened during the Flash Crash of May, 2010) and refrain from the arbitrage trade. Institutional investors then start to mistrust valuations across the board, high frequency traders no longer contribute to liquidity provision, and price divergence across trading venues worsen. The shock is transmitted through the network, and its effects are reinforced by a positive feedback. More shallow markets may increase the possibility of the feedback mechanism and therefore systemic risk, and vice versa. This line of research of WP4 will investigate this theoretical venues. We aim at providing a formal model for the occurrence of flash trading as triggered by liquidity shocks in markets populated by heterogeneous HFTs with different institutional characteristics and capital capacity. From the empirical perspective, we will focus on the spillover effects of flash crashes and on the time that the different markets take to recover, i.e. on market resilience, and on the causes of differences in “resilience” across the different markets.
3) The third line of WP4 focuses on accounting for interdependence among variables and the implications for systemic risk and, in particular, on the estimation of a financial networks with tools and methods commonly adopted in spatial statistics. In particular, the aim is to propose a new concept for measuring, forecasting, and managing systemic risk in presence of linkages among financial markets, monetary policy instruments, and sovereign debt in a coherent and encompassing setting. This line of research aims at developing an innovative and multidisciplinary approach to reinterpret these concepts. The main idea is to integrate econometric tools with network theory. This is a new methodological approach in this area. Models estimated using econometric methods are usually compared with the power of neural networks (Fioramanti, 2008) and networks are appreciated for their ability to model and forecast many non-linear and complex problems (Christakis, Barbaris and Spentzos, 2011).
 
This line of research works on the synergy between econometric estimations and network building up a toy procedure that shall be an added value also in other areas of research.