This work builds an agent-based model of financial markets to investigate how the interplay between high and low frequency traders can affect stock price volatility and the occurrence of flash crashes in financial markets. The increased occurrence and severity of flash crashes and excessive volatility of stock prices observed in financial time series have been associated to the rising importance of high frequency trading (HFT, see for instance, Sornette and Von der Becke, 2011, and further references therein). However, empirical and theoretical works have stressed both the benefits and the costs of HFT. On the one hand, earlier works have suggested that HFT improves market quality (Menkveld, 2011). First, liquidity has improved owing to the implementation of HFT in financial markets (Cartea and Penalva, 2011). Second, the development of HFT, which has changed the trading market structure, reduces transaction costs (Brogaard, 2010). Lastly, by strengthening the links between markets, HFT favors price discovery and market efficiency (Brogaard, 2010). On the other hand, earlier empirical and theoretical studies have highlighted possible concerns about the effect of HFT on market stability. In particular, the presence of HF traders seems to coincide with more frequent periods of illiquidity which is likely to explain the emergence of flash crashes (Paddrik, Hayes, Todd, Yang, Scherer, and Beling, 2011). Furthermore, HFT exacerbates market volatility (Zhang, 2010; Hanson, 2011; Kirilenko, Kyle, Samadi, and Tuzun, 2011) and affects negatively market efficiency (Wah and Wellman, 2012). However, so far the eﬀect of HFT on market quality is still controversial and no clear agreement has emerged on the actual impact of HFT on financial market instability.
This work contributes to the current debate on the impact of HFT on stock market dynamics by building an agent-based model wherein high frequency traders coexist with low frequency traders. More precisely, we aim at exploring which one among the commonly accepted characteristics of HFT, i.e., the speed or the sophistication of trading rules, can be a possible source of market instability. For this purpose, we extend the work by Westerhoff (2008) by introducing high frequency traders. More precisely, low frequency traders are of two types, fundamentalists and chartists. Fundamentalists set their strategies based on the fundamental value of the asset. Chartists set their strategies based on the observation of past price movements. Low frequency traders select one of these strategies every trading day according to the strategies’ past performance. Differently from low frequency traders, high frequency traders (HF traders henceforth) can trade many times within each trading day. In addition, HF traders do not accumulate large net positions. Lastly, HF traders do not hold overnight positions and try to close their positions by the end of the day.
Exploiting the above framework, we compare two scenarios based on these key commonly accepted features of HFT. In the first scenario, in line with earlier works, high frequency traders are zero-intelligent traders characterized by the very high speed of their order submissions (Kearns, Kulesza, and Nevmyvaka, 2010; Paddrik, Hayes, Todd, Yang, Scherer, and Beling, 2011). In the second scenario, high frequency traders are very sophisticated traders able to implement strategic decisions (see e.g. Easley, Lopez de Prado, and O’Hara, 2012). More precisely, besides their ability to trade at high frequency, HF traders also try to predict future price movements based on the observation of bid-ask spreads, volumes, past price movements and behavior of low frequency traders. In this case, HFT is modeled in a more sophisticated way in order to grasp the effect of their strategic behavior on market dynamics.
The main novelty of our study consists in the assumption of HF traders strategic behavior. In fact, to the best of our knowledge, among the few existing agent-based models for HFT, HFT is mainly modeled in a highly stylized way as zero-intelligence agents (Paddrik, Hayes, Todd, Yang, Scherer, and Beling, 2011; Aloud, Fasli, Tsang, Dupuis, and Olsen, 2012; Wah and Wellman, 2012). In contrast, we aim at accounting for the interplay of more sophisticated HF and LF trading strategies.
Our simulations show that, in presence of low frequency traders, high frequency traders are able to exploit market inefficiencies at a very low time-scale. In addition, HFT explain the emergence of flash crashes, stock prices excessive volatility and market instability. Furthermore, while high frequency traders are often characterized by their ability to trade at very high speed, our results suggest that HFT is particularly damaging because of the sophisticated strategies that such traders are able to use in the market. Lastly, we investigate the effect of the imposition of a financial transaction tax on high frequency trading and we examine whether this type of policy regulation can contribute to the stability of financial markets.
Aloud, M., M. Fasli, E. Tsang, A. Dupuis, and R. Olsen (2012): “Modelling the High-Frequency FX Market I: an Agent-Based Approach,” Discussion paper, Working Paper, University of Essex.
Brogaard, J. (2010): “High frequency trading and its impact on market quality,” Northwestern University Kel logg School of Management Working Paper. Cartea,
A., and J. Penalva (2011): “Where is the value in high frequency trading?,” Available at SSRN 1712765.
Easley, D., M. Lopez de Prado, and M. O’Hara (2012): “The volume clock: Insights into the high frequency paradigm,” Journal of Portfolio Management, forthcoming.
Hanson, T. A. (2011): “The Eﬀects of High Frequency Traders in a Simulated Market,” Available at SSRN 1918570.
Kearns, M., A. Kulesza, and Y. Nevmyvaka (2010): “Empirical limitations on high frequency trading profitability,” Available at SSRN 1678758.
Kirilenko, A., A. Kyle, M. Samadi, and T. Tuzun (2011): “The flash crash: The impact of high frequency trading on an electronic market,” Available at SSRN 1686004.
Menkveld, A. J. (2011): “High Frequency Trading and the New-Market Makers,” Discussion paper, Tinbergen Institute.
Paddrik, M. E., R. L. Hayes, A. Todd, S. Y. Yang, W. Scherer, and P. Beling (2011): “An Agent Based Model of the E-Mini S&P 500 and the Flash Crash,” Available at SSRN 1932152.
Sornette, D., and S. Von der Becke (2011): “Crashes and High Frequency Trading,” Swiss Finance Institute Research Paper, (11-63).
Wah, E., and M. Wellman (2012): “Latency Arbitrage, Market Fragmentation and Eﬃciency: An Agent-based Model,” University of Michigan Technical Report.
Westerhoff, F. H. (2008): “The use of agent-based financial market models to test the eﬀectiveness of regulatory policies,” Jahrbucher Fur Nationalokonomie Und Statistik, 228(2), 195.
Zhang, F. (2010): “High-frequency trading, stock volatility, and price discovery,” Available at SSRN 1691679.