Thirty Years Later?

In 1986, Olsen & Associates started collecting tick by tick Foreing Exchange (FX) data published by Reuters on their FXFX page. It was the starting of the high frequency adventure in Finance. Everything was new at the time, new methodology was needed to treat unevenly spaced time series, prices that came as bid and ask instead of one price. The data presented seasonality in volatility and not in the values themselves. We had to develop concepts like business time scales or intrinsic time to help model these time series. The sheer number of data was a challenge to filter in order to do meaningful statistics. It was a time for discovering scaling laws, fat tails, long memory and all those regularities that are now common knowledge.

Today 80% of the trades in the developed markets are due to high frequency traders. They use these data to build sophisticated algorithms for improving their business. In thirty years, the world of trading has completely changed, bringing up questions to wether they still play their roles for price discovery.

In this presentation, we will review the main advances that the high frequency data have brought to light and the approach that we developed over the years to understand the dynamics of financial markets. We will conclude on the recent advances and the questions that these new developments have brought to research both in market micro-structure as well as for macro-economics. We will also discuss the new risk landscape that has emerged from the advances of technology in finance.