City Research Online

Abstract

The aim of this thesis is to investigate the effects of foreign exchange order flows on exchange rate and stock market changes, in particular to examine the forecasting power of order flows and better understand the nature of the private information conveyed in order flows in the foreign exchange market. Chapter 1 investigates the performance of foreign exchange customer order flows (six major exchange rates over 3.5 years) as an additional explanatory variable to technical analysis to forecast exchange rate changes by applying genetic algorithm non-linear methodology. Using the interval permutations technique, we suggest that the improvement of order flows to the performance of technical analysis is not consistently present. Chapter 2 examines the role daily customer GBPUSD order flows play in explaining concurrent and future stock market changes in both UK and US, and discusses the heterogeneous effects from different groups of customers. The basic hypothesis tested is that if foreign exchange order flows have days-ahead effects on future stock market changes, it suggests that at least a part of the information carried by foreign exchange order flows is relevant for stock markets. Using daily GBPUSD order flows over 3.5 years from 2002 to 2006 provided by RBS, we find that: 1) impacts of order flows from corporate customers on stock markets are positive, while impacts of order flows from unleveraged financial institutions are negative; 2) impacts of corporate order flows are longer than those of financial order flows, especially for the US stock market, suggesting that the two groups of customers may hold different types of private price-relative information. We hypothesize that corporate customers of the bank are mainly based in the UK. When the world economy is doing well, multi-national companies are selling more goods in the US and repatriate more foreign currencies back to UK, during which more GBP or EUR are converted from US Dollars. More sales of US Dollars then reflect the good future prospects of the world economy and stocks listed in both US and UK will rise in value. For unleveraged financial institutions, when the world economy is going bad, clients of those mutual funds which are based in the UK will ask for redemptions of their funds. Assuming the bank services a client base that is UK oriented, this leads to the repatriation of money from abroad back to UK. The buying of GBP or EUR in exchange for US Dollars then takes place alongside sales of US and UK stocks. Foreign exchange flows into GBP or EUR from unleveraged funds forecast poor future stock market returns globally. Chapter 3 empirically tests the effects of EURUSD order flows from different groups of counterparties on the US stock market changes at high frequencies ranging from 1-minute to 30-minute, using a unique set of tick-by-tick order flows data obtained from a leading European commercial bank. We find that: 1) Order flows from “corporates” are positively related to exchange rate changes, while order flows from “financials” are negatively signed, which contradict many well-documented papers such as Evans and Lyons (2002a) (this high frequency forecasting power partly explain the failure of the trading strategy based on our daily order flows data in chapter); 2) The effects of order flows from “financials” are negative on stock market changes, while the effects of orders from “corporates” are positive on stock market changes, which further confirms our findings in chapter 2. Similar to chapter 2, the cross market effects documented in chapter 3 also suggest that there is information content in foreign exchange order flows and that it is likely to be macroeconomic in nature, relevant for stock markets. Chapter 4 concludes and suggests some directions of refinements and further research.

Publication Type:	Thesis (Doctoral)
Subjects:	H Social Sciences > HB Economic Theory
Departments:	Bayes Business School Doctoral Theses Bayes Business School > Bayes Business School Doctoral Theses

Export

Downloads