Stylized facts are the observed statistical properties of the traders’ trading behaviour in the financial markets.
Such stylized facts are established from the statistical analysis of financial market data. Gathering basic stylized facts with regard to the behaviour of traders and their trading activities is an important step in modelling real financial market traders.
In building a model of the financial market, stylized facts are considered to be the first verification criteria (LeBaron, 2006b). Without such facts, it is difficult to validate the accuracy of the agent-based models of traders.To establish stylized facts of the FX market traders’ behaviour, we need to explore high- frequency data related to their historical trading transactions in the market. High-frequency data allows us to explore some aspects of trader behaviour that could not be observed at lower frequencies (e.g. using daily data). Having surveyed the FX market traders’ behaviour literature, we find that previous studies have acquired a large sample of low frequency dataset of traders’ transactions, or a small sample of high-frequency datasets, neither of which are at account level. These do not allow a comprehensive study of actual FX market traders’ behaviour. Fortunately, our study is supplied with a high-frequency dataset, described in section 4.2, representing the physical transactions of over 40,000 anonymous accounts trading under the same terms and conditions.
In this section, we conduct a microscopic analysis of FX market traders’ trading behaviour to establish stylized facts about their trading behaviour. These stylized facts can be grouped under two main headings: correlation and the seasonality of trading activity. We find a daily and weekly periodic pattern of FX market trading activity. Additionally, correlation of FX market trading activity has been confirmed for the number of buy and sell orders, as well as for the number of opening and closing positions.
6.1.