| Summary: | 碩士 === 大同大學 === 資訊經營學系(所) === 107 === The financial calendar of a country is a time-course plan for the country to publish economic data or important financial activities so that citizens, businesses as well as stakeholders even in foreign countries could know the current economic state of the country. As far as currency trend is concerned, investors often use technical analysis to predict short-term price changes to determine how to invest, but technical analysis is based on the hypothesis of historical recurrence which is highly uncertain. Moreover, the trend of observation is often subjective, and the obtained indicators or price trends may not be perceived the same among observers, so the outcome is often diverse. In this paper we will study the relationship between economic events and currency development trend. A financial calendar often includes the data of the previous period and the forecast value of the current period. When the date/time is due, the data of this period is published. The fluctuation of foreign currency can be observed. As this is a prediction of the future, financial instruments like stock price, futures index, etc. will be affected, especially the currency exchange rate will be significantly influenced. Therefore, the amount of data published by the scheduled financial events at the time of expiration, and the extent of the impact on foreign exchange with high volatility, is worth studying. This study makes empirical research by comparing past data and comparing whether the data is related to exchange rate fluctuations. When the data is published and is consistent with the expected fluctuations in the exchange rate, the correlation between the two can be determined and used as a reference for investment decisions. This approach is a combination of basic, policy and news aspects. Since most of the past research focused on the impact of a country's domestic economic events, this study will extend to the economic growth and decline other countries. Here, we try to find the relationship between the US dollar and non-US currencies. We make forecast before the economic event is released, and the trend after the release is observed and discussed in depth. Since the impact of the news is mostly temporary, this study focuses on short-term fluctuations, that is, short-term movements after financial events or data releases. In general, the trend may be one of the following: conformation to expectations, contrary to expectations, or no significant change.
The research method used in this paper is an experimental one: we design a system whose functions are as follows: (1) draw data and events from the financial calendar; (2) analyze the data and the type of economic events, and based on the data and events to assess their impact and give prediction; (3) after the data or event is published, the outcome is compared with the expected value and give it with a score; (4) The score of the latest data or event is aggregated to the current time period. (5) Re-assess the strength of the currency based on the newly accumulated score.
The application of this system is to support foreign exchange margin trading. Due to the large leverage of margin trading and the sensitivity to volatility, our trading strategy is focused on the very short-term trend; that is, the trend of foreign exchange within one hour after the announcement of financial data/events. If the trend is obviously strong, it may be the beginning of a long-term trend, which can also provide a basis for decision-making for other applications (such as hedging). To understand the availability of this system, we use simulation to test the applicability of the system. Prediction made by the system is mostly correct, matching with the actual trend. Moving span is 90 days. From September 2018 on, we test the system with actual market data: 56 orders were placed and 39 of them won and 17 lost, achieving profit rate of 70%. The initial capital of USD 550.-- gained $85.--, implying the applicability of the system.
|
|---|
