Algorithmic trading

How does algorithmic trading affect financial markets?

Much of the current public debate about computer-based trading or algorithmic trading (AT) is concerned with the class of aggressive predatory algorithms, especially those that operate at high speed and with high frequency. The raison d’être for financial markets is to aggregate myriad individual decisions and to facilitate an efficient allocation of resources in both primary and secondary markets by enabling a timely and reliable reaping of mutual gains from trade, as well as allowing investors to diversify their holdings. As with many other aspects of modern life, innovations in technology and in finance allow the repetitive and numerically intensive tasks to be increasingly automated and delegated to computers.

Automation, and the resulting gains in efficiency and time, can lead to benefits but can lead also to private and social costs. The focus of our study is solely on possible repercussions of AT (including high frequency trading (HFT) in particular) on stability, and convergence of markets to equilibrium.

Surprisingly, there have been very few experimental studies of the interaction between humans and robots in financial markets. Policy and regulation are at present dictated by theoretical arguments (is “model based”), or by statistical analysis of historical data comparing financial markets features across epochs with and without robots. Within the set of the few experimental studies, ours stands out in that we allow humans to control the robots – humans not only design them, but also switch them on and off.

In a first stage, we have been observing price formation and trade in a canonical multi-period asset pricing setting that is known to robustly generate bubbles and crashes, the so-called Smith-Suchanek-Williams setting. Traders can trade manually or upload and launch one from a large variety of robots, and even trade manually alongside their chosen robot. Initial evidence seems to indicate that robots make markets more efficient in the sense of Fama (less autocorrelation in price changes, tighter bid-ask spreads) but we do not see any fewer bubbles…

This line of research makes heavy use of the online markets software we developed, namely, Flex-E-Markets.

The research is done in collaboration with Elena Asparouhova and her lab ULEEF at the University of Utah.