Olsen & Associates
Olsen & Associates (commonly known as O&A) was founded in Zurich in 1985 by Richard Olsen as a think tank in finance. In the mid-1980s, as computers were becoming faster and computer memory more affordable, it became possible to record and study intraday tick-by-tick financial data. Hitherto, almost all research in modeling market microstructure was done using daily data. The vision of O&A (and its founder) was to take early advantage of the digital revolution (starting in the mid-1980s) and use high frequency intraday data along with ideas and concepts from mathematics and physics to model financial markets. This explains why the O&A research team[1] largely consisted of PhDs from those fields. O&A funded its research by using its research results in financial consultancy and its real-time forecasting and trading signal services called the Olsen Information System (OIS).[2][3]
Research
The foreign exchange market is well known as one of the fastest-moving markets generating the most intraday ticks. It is no surprise, therefore, that O&A first started collecting intraday data for this market and consequently the first comprehensive suite of worldwide research into intraday market microstructure focuses on this market.
One key insight of O&A (deepened by discussions[4] with Benoit Mandelbrot) was that the use of intraday data would reveal a fractal nature of markets.[5] This work culminated in the discovery of the scaling law.[6][7] Another notable contribution was the work of the O&A research team with 2003 Nobel prize winning economist Robert F. Engle in modeling intraday market data using (ARCH family) stochastic volatility models.[8][9]
A compilation of all significant research by the O&A research team: *An Introduction to High-Frequency Finance*[10] was published in 2001.
Conferences
Leveraging its position as a pioneer in developing research and interest in the area of high frequency intraday data, O&A hosted the first international conference on High Frequency Data in Finance (HFDF-I) in 1995, followed by a second conference (HFDF-II) in 1998. Associated with each conference, O&A released unique high frequency intraday data sets, HFDF93 and HFDF96, respectively, with the intent that research presented at the conferences would be based on these standard data sets. Subsequently, another data set, HFDF2000, was also released to encourage even more academic work worldwide on a common data set.[11]
Mathematical techniques
As mentioned before, prior to the path-breaking O&A initiative to capture and study intraday data, research in economics was primarily done with daily data. By its very nature, daily closing data consists of evenly spaced time series (ignoring market holidays). On the other hand, intraday tick data consists of unevenly spaced time series because ticks can appear at any time. It was, therefore, incumbent on the O&A research team to formalize methodology for handling such series. It is well known that such series can be converted into evenly spaced series using linear or previous tick interpolation methods; however, such methods are not sensitive to all the available information regarding the time lapse between ticks. O&A, therefore, introduced the set of EMA (Exponential Moving Average) operators[12] to handle the creation of indicators from intraday data. Not only are these techniques mathematically superior, leading to smoother indicators, but they also lend themselves to more elegant programming, as a single update handles the attenuation of older price changes compared to rectangular moving averages where older contributions to the average need to be explicitly removed when adding new data.
Another very important innovation of O&A (inspired by insights from its founder) was the concept of a de-seasonalizing time scale (referred to as Theta time in O&A jargon).[13] In this time scale, the average intraday market activity, which predictably varies within the day (for example, peaking when North American market makers wake up and become active in the European afternoon, or in the interesting case of dips like the Japanese lunch hour impacting Japanese Yen markets), appears to become flat. In other words, by analyzing the daily routine of market activity averaged over long horizons, a new time scale is constructed such that time stretches when markets are predictably inactive and compresses when they are predictably active.
Software
In order to systematically research high frequency intraday data, O&A had to first start recording live data from financial data services in the late 1980s, like Reuters, Telerate, and Knight Ridder. In those early days, data updates from these services arrived over telephone lines with single character/digit updates referring to a grid location on the display. The nature of the data transmissions was to satisfy a human watching a live screen and was not tailored for recording data. O&A had to deal with this caveat when developing its early data collection software. It was only later, in the early 1990s, after recognition of the value of recording intraday data and the growth of the internet, that transmission methods became more conducive to non-display recording of live data.
Further, O&A developed a proprietary time-series database system which supported capture of live high-frequency data while simultaneously supporting queries of live or historical data for client applications.[14]
As stated earlier, O&A first focused on the foreign exchange market, which, being an over-the-counter market, meant that quotes would originate from several hundred contributors (market makers and brokers). Not all contributors would offer serious quotes in line with existing spreads, and occasionally there would be gross outliers due to transmission or human errors on the contributor side. To handle this issue, O&A developed special real-time filter technology[15] to be able to convert the live raw data into research-quality data.
Olsen Data
After the publication of the first set of papers using intraday data by the O&A research team and the several papers published after HFDF-I and HFDF-II and the release of the standard data sets HFDF93, HFDF96, and HFDF2000 referenced in those papers, O&A began to receive requests for intraday tick data from worldwide academic institutions who wanted to publish their own results using intraday data as well as banks, hedge funds, and investment houses doing private research. O&A dedicated a business unit to this service called: Olsen Data. Eventually, Olsen Financial Technologies (OFT)[16] was founded as a separate company which inherited all the O&A-developed technology and the database of high-frequency data which starts from February 1986. OFT continues to collect high-frequency data and runs today as a small operation supplying financial data and consultancy services in market risk management and data management.
Olsen spinoffs
O&A suffered a financial setback in 2001 and had to be shut down. The O&A research group disbanded. A holding company, Olsen Ltd, continued with the intent of being an incubator for spinoff companies using Olsen Technology, some of which are listed here.[17] Not all of these companies survive today. Prominent among the surviving companies are Oanda[18] (which sounds like O&A) and Olsen Financial Technologies[16] (Olsen Data).
References
- ↑ "Olsen & Associates: Research". web.archive.org. 1997-06-16. Archived from the original on 1997-06-16. Retrieved 2021-03-20.CS1 maint: Unfit url (link)
- ↑ Name (1992-10-19). "OLSEN OF ZURICH AIMS TO CUT RISK IN CURRENCY TRADING". Tech Monitor. Retrieved 2021-03-31.
- ↑ Flint, James (July 1996). "The Dynamics of Capitalism". Wired UK. No. 2.06.
- ↑ Muldoon, Oliver (2019-10-14). "The Wandering Scientist Turned Father of Fractals". Medium. Retrieved 2021-03-19.
- ↑ Davidson, Clive (Dec 15, 1997). "WILDLY RANDOM MARKET MOVES". Journal of Commerce. Archived from the original on July 11, 2021. Retrieved July 5, 2021.
- ↑ "Statistical study of foreign exchange rates, empirical evidence of a price change scaling law, and intraday analysis". Journal of Banking & Finance. 14: 1189-1208.
- ↑ "Scaling Law" (PDF). citeseerx.ist.psu.edu. Retrieved 2021-03-20. Unknown parameter
|url-status=ignored (help) - ↑ "Timeline" (PDF). Unknown parameter
|url-status=ignored (help) - ↑ "About Olsen - Leadership in Forecasting Technology: An Olsen Timeline". web.archive.org. 2007-10-31. Archived from the original on 2007-10-31. Retrieved 2021-03-20.CS1 maint: Unfit url (link)
- ↑ An Introduction to High-Frequency Finance. ISBN 978-0122796715. Search this book on
- ↑ "HFDF Data Sets" (PDF). Unknown parameter
|url-status=ignored (help) - ↑ "EMA operator" (PDF). Unknown parameter
|url-status=ignored (help) - ↑ "Theta Time" (PDF). Unknown parameter
|url-status=ignored (help) - ↑ "Olsen Data Repository" (PDF). Unknown parameter
|url-status=ignored (help) - ↑ "Olsen Data Filter" (PDF). Unknown parameter
|url-status=ignored (help) - ↑ 16.0 16.1 "OlsenData: Welcome". www.olsendata.com. Retrieved 2021-03-20.
- ↑ "Olsen - Advancing the Frontiers of Finance". web.archive.org. 2006-04-27. Archived from the original on 2006-04-27. Retrieved 2021-03-20.CS1 maint: Unfit url (link)
- ↑ "Our History | OANDA Up To Now | OANDA". www.oanda.com. Retrieved 2021-03-20.
External links
This article "Olsen & Associates" is from Wikipedia. The list of its authors can be seen in its historical and/or the page Edithistory:Olsen & Associates. Articles copied from Draft Namespace on Wikipedia could be seen on the Draft Namespace of Wikipedia and not main one.
