Download as pdf or txt
Download as pdf or txt
You are on page 1of 18

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/346037586

ARTIFICIAL INTELLIGENCE (AI) IN THE FOREIGN EXCHANGE MARKET

Chapter · November 2020

CITATIONS READS
0 351

1 author:

Tarana Azimova
Istanbul Aydin University
12 PUBLICATIONS   3 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Tarana Azimova on 20 November 2020.

The user has requested enhancement of the downloaded file.


In: Emerging Human and Techno-Human … ISBN: 978-1-53618-602-4
Editors: Christina Koutra et al. © 2020 Nova Science Publishers, Inc.

Chapter 5

ARTIFICIAL INTELLIGENCE (AI) IN THE


FOREIGN EXCHANGE MARKET

Tarana Azimova*, PhD


Business Administration, Istanbul Aydın University, Istanbul, Turkey

ABSTRACT

The foreign exchange market- the world’s largest, dynamic and


intricate financial structure has been in pursuit of technological
advancement since it came into existence. With the passage of time, data
analysis, communication, market monitoring, decision-making and trading
in the foreign exchange have become increasingly computerized and
electronic. The foreign exchange market has undergone three major
changes in terms of technological innovation. The first technological
innovation in financial markets was the introduction of electronic trading;
the second one was the Internet; and the third and present one is the
introduction of Artificial Intelligence (AI), mainly used for exchange rates
prediction. The application of AI, in turn, which has increased significantly
in the last few decades, enabled market participants to improve
communication, their accessibility to the market, and an overall market
efficiency.

*
Corresponding Author’s Email: tarana.azimova@nisantasi.edu.tr.
82 Tarana Azimova

Keywords: foreign exchange market, artificial intelligence,


communication ability, market efficiency

INTRODUCTION

The foreign exchange market is one of the largest financial markets in


the world operating in a highly complex trading environment. The first
online foreign exchange trading platform was launched in 1996; this allowed
retail traders to take part in the market (Cartea, et al., 2019; Violeta, 2010).
However, since then, it has significantly evolved; mainly assisted by the
introduction of progressive technology aimed at deploying AI devices.
Additionally, since its commencement, the foreign exchange market has
been explicitly designed to promote new trading experience, skills and
knowledge; it is a unique market, where multiple advancing technologies are
combined to establish a modern trading platform.
Various innovations in trading technology, such as online trading
platforms, mobile trading and hybrid AI systems, among others, have had a
generally positive impact on the market-trading turnover at different stages,
as discussed below. For example, in 1996, the average global daily turnover
of the online trading platform was US$1,633 Billion. Since its
commencement, the foreign exchange market was perceived as being an
expanding market. The only downfall was in 2001, when the average daily
turnover plummeted to US$1,705 Billion. Two decades later, the average
daily turnover in the foreign exchange market amounted to US$6,514 Billion
(BIS 2018).
The foreign exchange market has gone through three main stages of
technological changes. The first stage, the “technological frontier one”,
refers to electronic trading which commenced when Reuters introduced the
Reuters Data Market Services in 1981 (Esen, 2018; Rime, 2003), and was
used to communicate trading interests between market participants.
The second stage, the “technological frontier two”, refers to the early
years of the art technologies’ revolution such as the Internet. Internet trading
is comparatively new in the foreign exchange market. The first non-bank
Artificial Intelligence (AI) in the Foreign Exchange Market 83

Internet site for customer trading Deal4Free, was developed in the USA in
1996 (Esen, 2018; Rime, 2003). The Internet served as a significant catalyst
in the effort to increase communication and trade between nations. New
search engines such as Google changed the complete face of international
trade and increased the demand for the foreign exchange market as traders
could effectively communicate around the world, set orders on the spot,
forward and swap exchange markets. The Internet has in effect demolished
all boundaries around countries and unified the exchange markets (BIS,
2018). The foreign exchange market currently offers new trading
opportunities through mobile solutions while foreign exchange market
participants use mobile applications that enable them to invest, or open
positions in the market, close positions, and track their profits
instantaneously and at a very low cost.
Nowadays, the foreign exchange market is moving towards what could
be defined as a “technological frontier three”, which refers to the
introduction of even more complex technologies, such as hybrid AI systems,
used to predict exchange rates.
Accordingly, this chapter discusses the way AI trading affects the
foreign exchange market. Specifically, it discusses the phenomena arising
out of using AI in the foreign exchange market; the implications of emerging
technologies on the way the participants communicate and get access to the
market; and the impact of new technologies on market efficiency. An
attempt is made to to fill a conceptual gap and contribute to the existing
academic literature on recent technological advancements and the increasing
speed of adopting new technologies in the foreign exchange market.

THE IMPACT OF ARTIFICIAL INTELLIGENCE TRADING


ON THE FOREIGN EXCHANGE MARKET

This section provides information on the way technological and AI


trading influences foreign exchange markets. It discusses several
dimensions of the positive impact of technological advancement on trade on
84 Tarana Azimova

forex including communication ability, market accessibility and market


efficiency.

Communication in the Foreign Exchange Market

Most of the transactions in the foreign exchange are carried out via
interbank, which is the primary money market. Because inter-bank
transactions represent over 90 percent of the total volume of the foreign
exchange market, and have no centralized physical market place,
participants’ interaction is assisted by technology (BIS, 2018).
The foreign exchange market may be depicted as follows: (i) the market
is geographically dispersed and does not have a centralized market system;
(ii) there is uninterrupted trading across the markets; (iii) there is a great
number of market participants that provide liquidity to the market; and (iv)
the trading volume is big and increases rapidly over time (Tygier, 1986)
Therefore, having these characteristics, the foreign exchange market
heavily relies on technology, which influences the way foreign exchange
participants, such as large commercial banks, businesses, foreign exchange
brokers, and central banks interact with each other.
From a historical perspective, in older days, trading in foreign exchange,
could be described as primitive, as the number of transactions and the
participating countries and companies were limited (Tygier, 1986).
However, in the last few decades, foreign exchange has experienced major
changes. Starting in the 1970s, direct trading was performed by telex or
telephone. From 1981 onwards, the foreign exchange started to apply the
Reuters Data Market Services for communicating trading interests; first
introduced by Reuters. The system was a closed network for bilateral
electronic communication. Even though the system was considered as
technologically advanced, the Reuters Data Market Services did not become
an evolutionary system for the foreign exchange market (Esen, 2018; Rime,
2003).
It was only in 2000 that the conventional trading technology gradually
started to re-engineer and become the dominant trading and interaction
Artificial Intelligence (AI) in the Foreign Exchange Market 85

Reuters D2000-1 communication system (Esen, 2018; Rime, 2003). For


instance, a market maker or a dealer is an active trader in the market. He/she
contacts other dealers for price quotations, offer bids and ask prices. Dealers
conduct direct trade over a telephone or computer terminals and mostly
specialize in one currency. To fulfill the transaction, the dealer has a
provision of a particular currency. The stockpile of a particular currency at
the same time defines the limitations of a dealer’s market activity. Market
makers interact on price quotations through the conversational intelligent
terminal, named Reuters D2000-1 communication system (Rime, 2003).
Reuters developed a new subscriber line protocol to augment the
efficiency of the traders’ connection to the data center especially over poor
quality satellites. In addition, the system dramatically increased the network
capacity by introducing nodes connections. As a result, dealers can use the
terminal to establish several simultaneous conversations. Figure 1 displays
communication example over Reuters D2000-1(Rime, 2003).

Source: Rime, (2003). New Electronic Trading Systems in Foreign Exchange Markets.

Figure 1. Interdealer communication on Reuters D2000-1.

The first and the second lines in Figure 1 show that the order came from
the counterpart who is the dealer. The order shows the name of the dealer,
institution code, date and number assigned to the order. The third line
86 Tarana Azimova

provides information on the order; it is evident that there is a demand for a


spot DEM/USD quote for up to $1 million. Line four provides information
on the actual bid made and asks for quotations. For example, the bidding
quote is 1.8145 and the asking quote is 1.8147. If the counterparty agrees to
the transaction, then it will buy$1 million at the bidding quote of 1.8147.
The transaction price always equals the asking quote for the counterparty.
Once the transaction is over, the record verifies specific details on the
transaction such as the exact price, quantity and valuation date. The
conversation ends up with traditional phrases such as “thanks for deal”,
“thanks and goodbye”.
Technological advances that have changed the face of trading across a
range of foreign exchange markets include lower storage costs, greater
processing power, and the ability to transfer information almost instantly.
This progress has enabled participants to respond to developments in near-
real time, and have opened the doors to deeper analysis, significant
efficiency gains and new business models (BIS, 2018).
Presently, foreign exchange markets use various innovative means of
communications such as advanced telephones and applications, computer
terminals, fax, telexes, and online currency trading platforms amongst
others. Modern age technology and technological diversity creates
opportunity for foreign exchange market participants to establish instant
communication at a very low cost. Therefore, technology is facilitating the
flexibility and effectiveness of interactions.

Trading and Market Accessibility

Currently, advanced technology allows foreign exchange markets to


move towards higher levels of globalization. It is a good example of how AI
in fact breaks down all geopolitical boundaries, and makes different market
segments accessible and interconnected. A decade ago, customers’ access to
the foreign exchange market activity was limited. Customers depended on
banks for almost all financial services. They requested quotes from banks
over the fax and telephone. The banks’ representatives would carry out
Artificial Intelligence (AI) in the Foreign Exchange Market 87

financial services with certain commissions. Since then, the new technology
decreased dramatically the extent to which customers relied on banks
(Dmour, 2019; Choi, 2000).
In 1981, Reuters developed the first electronic trading system, Reuters
Dealing 2000-1, which was similar to a network supporting bilateral trade.
However, Internet trading is comparatively new in the foreign exchange
market. The first nonbank Internet site for customer trading Deal4Free, was
developed in the USA in 1996. The introduction of several nonbank Internet
sites such as IFX Markets and MatchbookFX in 1999, HotSpotFX and
OANDA in 2001, and ChoiceFX in 2002, allowed market participants in the
USA to start customer-to-customer trading (Esen, 2018; Rime, 2003).
According to Rime (2003), most of these systems are organized as
crossing networks, but they try to resemble electronic brokers. Crossing
networks are trading systems that obtain their prices from another trading
venue; hence there is no price discovery (price discovery is the process of
determining the actual price of an asset). However, systems like the
Currency Management Corporation (CMC)’s Deal4Free resemble the
traditional direct trading and it has its own price discovery. Other systems
like ChoiceFX depend on limit orders from customers (a limit order is an
order that is placed by an electronic broker to trade a security at or better
than a specified price).
Initially, a banks’ response to a nonbank Internet trading platform for
customers was to install their own customer sites. Nowadays, however,
almost all transactions in the over-the-counter market are carried out via the
interbank market (global network used by financial institutions for trade).
Modern time technology, such as online currency trading platforms and the
internet offer a gateway to the interbank market. Regarding the bank–
customer relationship, Joel (2017) suggests that the Internet trading can
become so successful that banks could lose their entire customer trading.
Presently, the foreign exchange market is still a hybrid market in the sense
that it combines Internet and conventional trading (Joel, 2017).
One significant outcome in the use of AI in foreign exchange markets
due to market accessibility is an increase in the market’s liquidity, and hence
an increase in the daily turnover (see Table 1).
88 Tarana Azimova

Table 1 demonstrates that the daily turnover in the biggest Over the
Counter (OTC) foreign exchange markets has expanded overall. In 1995 and
1998, the average global daily turnover of the online trading platform was
$1,633 Billion and $2,099 Billion respectively. In 2010, the daily turnover
increased by almost 3 times, when compared to 2001 and amounted to
$5,045 Billion. The only downfall was in 2001, when the average daily
turnover decreased to $1,705 Billion. In 2016, the average daily turnover in
the foreign exchange market increased by almost 30 percent compared to
2015 and amounted to 6,514 Billion US Dollars. The statistics illustrate that
AI allows foreign exchange to be less manipulative and almost unbreakable.
Fierce competition made technology development companies reduce the
price in the use of technology for foreign exchange markets. This in turn
made it more accessible and available to not only large entities with big
funding sources, but also to smaller institutions and entities. By offering a
gateway to the main money market, technology has allowed smaller entities
to have an equal access to similar level liquidity in the OTC market.

Table 1. Daily trading turnover of (over-the-counter) foreign


exchange, by country April 1995–2016 daily averages,
in billions of US dollars

1995 1998 2001 2010 2016


United Kingdom 479 685 542 1,854 2,406
United States 266 383 273 904 1,272
Japan 168 146 153 312 399
Singapore 107 145 104 266 517
Hong Kong 91 80 68 238 437
Total 1,633 2,099 1,705 5,045 6,514
Source: BIS, 2018

Market Efficiency

The concept of market efficiency, which was first developed by Eugene


Fama in 1970, states that it is not possible for an individual investor to beat
Artificial Intelligence (AI) in the Foreign Exchange Market 89

or outperform the market. According to the efficient market hypothesis,


market anomalies will not exist in the market, because the arbitrage activity
will eliminate all price differences between the markets (Fama, 1970).
Arguably, a well-functioning market is linked to the availability of
information. A broad base of investors can process this information to make
investment decisions that may form the future of financial markets.
Therefore, efficiency in the foreign exchange market depends primarily on
the number of market participants, the availability of accurate information,
and the openness of the market globally.
Technological advancement supports instantaneous flow of information
in financial markets. For example, once traders adjust their orders in one
venue, this information is reflected almost instantaneously on other markets.
The pace at which the market information becomes available has also
increased. Trading platforms such as Reuters Matching and Electronic
Broking Services increase the frequency of pricing updates. For example,
the frequency of data updates on the Electronic Broking Services increased
from every 100 milliseconds (ms) to every 20ms in September 2016 (BIS,
2018). This increase in data update has had a positive impact on activity in
foreign exchange markets.
AI allows dissemination of information to a broad base of participants
at low cost and in real time. Therefore, expanding the universe of market
participants with access to information. Alexis (2018) and Choi (2000) argue
the introduction of web technology has doubled the frequency of trading.
The most important contribution of technology to the foreign exchange
market, as mentioned above, is the reduction in trading costs. Hence, any
market participant who has Internet connection can have constant access to
news, historical and real time return data, macro and micro economic
statistics, listed companies financial reporting data, analysts’ expectations
and forecasts, and updates on major changes in the business world among
others.
A significant result is that technology has increased levels of
competition in the foreign exchange market. Because of intensive
competition and heavier trading in the markets individual traders now make
considerably lower profits than a decade ago. Moreover, technology
90 Tarana Azimova

supports traders with uninterrupted updates on the performance of their


portfolios. The amount of information that the investors currently own is
overwhelming when compared to just ten years ago.
Foreign exchange markets are globally decentralized, having no
incorporated central market mechanism. Because trading in the markets is
conducted worldwide, it is extremely difficult to observe the price
formations in the markets. However, technological progress allows traders
to observe markets and price formations even for minor intervals. For
example, whenever price difference occurs for the same currency pairs in
different markets, the trader using internet trading will immediately notice it
and will take action to make a profit. The pricing errors may last for several
seconds; therefore, it is important for the trader to take quick action.
Technology in this case is used to support simultaneous purchase and sale
actions. The trader can purchase the currency at a lower price in one market
and simultaneously sell the currency in another market at a higher price,
hence making a quick profit.
Arbitrage maintains a unique mechanism to guarantee that the
quotations do not diverge from their fair value. Arbitrage exists because of
market distortions or inefficiencies and therefore is a vital activity to
maintain the efficiency of the financial markets. With technological
advancement, it has become very easy for arbitragers to observe the
quotations around the world and to profit from pricing errors in foreign
exchange markets. Hence, it could be deduced that AI supports market
efficiency by reducing boundaries and accelerating trade around the world.
The next section provides further information on the actual state of the
foreign exchange market in terms of application of AI devices to provide
exchange rate forecasts.

DECISION MAKING AND FORECASTING IN FOREIGN


EXCHANGE WITH AI DEVICES

The foreign exchange market is one of the world’s most popular


investment platforms, whilst the environment it takes place in is rather
Artificial Intelligence (AI) in the Foreign Exchange Market 91

complex. Different factors can influence exchange rates movements. As


these factors are both qualitative and quantitative in nature, modeling some
factors can be straightforward, while others are difficult to quantify. AI
devices would normally allow investors to model both quantitative and
qualitative factors. Quantitative factors that are outcomes of macroeconomic
and microeconomic changes are measurable in numeric terms. For instance,
change in trade balance conditions or interest rates are quantitative
measures. Qualitative factors on the other hand are more difficult to measure
but these factors have an impact on quotations. Global economic events,
country specific economic events, political situation, exchange rate
speculations, trader’s psychological expectations, and trader’s psychological
behavior among others are examples of qualitative measures.
Academic literature shows that fluctuations in returns are not random
and that returns move in a highly dynamic and non-linear manner (Tsaih,
1998). To bring tangible solution to foreign exchange trading decisions,
academic literature suggests the implementation of hybrid AI systems.
Academic investigations apply neural networks to predict foreign exchange
price in terms of quantitative data, and an expert system AI system to handle
qualitative parameters (Paris, 2019; Yu, 2005).
A neural network is a number of algorithms used to diagnose primary
relationships between parameters. These AI applications comprise layers of
interconnected nodes, which operate similar to the neural network of the
human brain. A neural network, which gains popularity in the currency
trading and interpretation of exchange rate patterns, has several advantages
such as robustness, conformity and nonlinearity. Another AI system
popularly used to improve the decision quality and cut the costs of
consulting services in the foreign exchange is an ‘expert system’. This Al
technique is in fact a computer-based system, which demands high-quality
information and is used to understand complex interactions and relationships
between parameters. The main advantage of AI systems as compared to a
human expert is their capacity to solve complex tasks, while working
continuously without getting emotional and tedious (Paris, 2019).
Because the foreign exchange market is a highly intricate environment
and a number of parameters influence exchange rates, studies use both
92 Tarana Azimova

quantitative and qualitative analysis. Miikkulainen et al. (2019) suggests that


AI techniques help to observe all tendencies on real time basis. On the other
hand, sophisticated technological tools make it easier to handle big data, and
understand and compare different parameters using both historical and real
time data. In addition, AI makes it possible to see the trends in different
variables and to understand interactions of different parameters. As a result,
advanced technologies support trading decisions and increase the speed of
the transactions (Miikkulainen et al., 2019).
Some empirical studies such as Alexandridis et al. (2018), Yu (2005),
Nedovic (2002), Guimaraes (1994) found that the neural networks and
expert systems combined together have increased the accuracy of
forecasting and have performed well. However, these researches do not
distinguish between the long term and short-term periods. AI algorithms can
make accurate short-term predictions as the investigations suggest, but they
are also reliable for making long-term forecasts. The challenge is that there
is more ambiguity associated with long-term forecasting. More events can
occur or the present events may no longer trigger the quotation changes in
the future. That is why different number of parameters need to be
considered. If assumed that certain number of parameters stay constant for
one week, then modern AI applications can be applied to forecast currency
prices over a period of one week.
However, because most of the variables are often unpredictable and
subject to frequent change, it is impossible to make predictions over a longer
period. That is why traders are more likely to rely on derivatives to manage
their portfolio risk for longer periods. Additionally, instead of relying on
modern predictive analytics tools, traders in foreign exchange markets
possibly hedge, or protect their currency positions using financial
derivatives such as forward, futures, options and swaps.
On the other hand, AI techniques do not provide a creative response in
unusual circumstances. Because AI systems are sets of algorithms, they can
provide solutions based only on the information fed to the system. Given
that the foreign exchange market is a type of financial market based on
international trade, resulting in an erratic system with no central governing
body, analysts try to acquire data from a variety of trades among
Artificial Intelligence (AI) in the Foreign Exchange Market 93

international currencies. However, not all of the requisite data can be


acquired or incorporated into the system. In such cases, the data in question
is totally omitted from the analysis. Because the system totally depends on
objective data, inadequate knowledge or errors in data can lead to wrong
decisions. Therefore, to improve decision quality and forecast reliability,
foreign exchange markets currently, rely on both conventional and Al
systems.
Foreign exchange markets’ participants argue segments of the financial
system are employing AI and machine learning only to a certain extent.
Whilst AI is increasingly applied in other areas, such as fraud detection,
capital optimization, and portfolio management applications. Liebergen
(2017) suggests that companies use AI for fraud detection, credit monitoring
and risk mitigation services at financial institutions. Struntz (2017) affirms
that the insurance industry is increasingly adopting machine learning to
analyse complex data. In the United States, United Kingdom, Germany and
China, the insurance industry has adopted AI and machine learning
applications. According to the report prepared by the Fınancial Stability
Board (2017), however, there is limited information on the extent of
adoption of AI in various markets and segments and futher investigation is
needed. Therefore, it is most likely that modern predictive analytics tools are
not adequately developed to suit the foreign exchange market.

CONCLUSION

This chapter discusses implications of the growing use of AI in foreign


exchange markets. AI made its way into the foreign exchange to handle the
complicated processes happening in the marketplace. AI uses sophisticated
statistical and mathematical models to assist the most actively traded
markets in the world. The foreign exchange market is a good example of
how AI in fact eliminates all geopolitical boundaries and makes various
market divisions accessible and interconnected.
Further, it illustrates that AI influences foreign exchange markets in
different ways, including but not limited to communication efficiency,
94 Tarana Azimova

accessibility of the market, market efficiency and decision-making. The use


of AI is improving communication among foreign exchange participants
such as large commercial banks, businesses, foreign exchange brokers, and
central banks. Because inter-bank transactions represent over 90 percent of
the total volume of the foreign exchange market, and it has no centralized
physical market place, participants in the market interact with the help of
technology. Transactions in the market take place over the telephone, fax,
telexes, computer terminals, and other means of communication. Therefore,
assisted by modern age technology, market participants can enjoy
communication flexibility and therefore interact more effectively.
Conversely, technology has dramatically increased accessibility to the
foreign exchange market. At the same time the fierce competition between
companies allows market participants to use technology at reasonable prices.
As a result, not only the foreign exchange market is available to large entities
with big funding sources but it is also becoming available to the smaller
institutions and entities. By offering a gateway to the main money market,
technology allows smaller entities to have an equal access to the similar level
of liquidity in the OTC market. In addition, the higher the access to the
market using technology, the greater a trader’s capacity to understand what
is happening in the marketplace and the quicker and more efficiently they
process big volume data.
Technological advancement has had a positive influence on the market
efficiency. There are some key requirements for increasing the foreign
exchange market efficiency- the availability of a broad base of participants,
the availability of accurate information, and the openness of the market
globally. AI is rapidly democratizing foreign exchange markets, as it
increases the visibility of the markets, and reduces the boundaries to enter
the market, thus resulting in higher competition, which in turn decreases
transaction costs. Costs of carrying out trades are also being reduced to
insignificant levels, while processes enable different income level groups to
take part in the financial system. Moreover, AI allows information
dissemination to a broad base of participants at low cost and in real time,
therefore expanding the universe of market participants with access to
information.
Artificial Intelligence (AI) in the Foreign Exchange Market 95

Further, AI promotes decision-making process in the market, while AI


devices allow investors to observe all tendencies in real time, fulfill real time
analysis, and make an accurate assessment on market events. Sophisticated
technological tools make it easy to understand and compare different
parameters using both historical and real time data. In addition, AI makes it
possible to use vast amounts of historic data and observe the trends in
different variables, and to understand interactions of different parameters.
The foreign exchange market is now undergoing the process of adoption
of AI systems in the field of exchange rate determination and forecasting.
Exchange rate volatility is considered a barrier to trade, as it increases the
risk level of the financial markets. Even though a lot of research was
undertaken to predict exchange rates, it was concluded that exchange rates
are extremely difficult to estimate. Some companies like Nikkei, Altredo and
ROFX claim that they use AI to make precise forecasting of the future value
of currencies. However, little is known about the software and methods they
apply to handle this challenging task. It is most likely that the development
of modern predictive analytics tools for exchange rate determination and
forecasting is not taking place in the forex space now.

REFERENCES

Alexandridis, A. K., Panopoulou, E. & Souropanis, I. (2018). Forecasting


Exchange Rates: An Iterated Combination Constrained Predictor
Approach, Working Paper No. 003.
Bank for International Settlement. (2018). Monitoring the Fast Paced
Electronic Markets. ISBN 978-92-9259-193-9. Retreıved from:
https://www.bis.org/statistics/d11_2.pdf.
Cartea, A., Jaimunqal, S. & Walton, J. (2019). Foreign Exchange Markets
with Last Look, Mathematics and Financial Economics, 13(1), 1-30.
Esen, G. (2018). Double Light Speed: History, Confusion, and Recent
Applications to High Speed Trading, Norwegian University of Life
Sciences.
96 Tarana Azimova

Fama, F. E. (1970). Efficient Capital Markets: A Review of Theory and


Empirical Work, Journal of Finance., 25(2), 383-417.
Financial Stability Board (2017). Artificial Intelligence and Machine
Learning in Financial Services. Market Developments and Financial
Stability Implications.
Gaucan, V. (2010). Introduction to the Foreign Exchange Market. Journal
of Knowledge Management, Economics and Information Technology.
MPRA Paper No. 28078.
Gene, D., Efi, G. & Andrei, S. (2001). Technology, Information Production,
and Market Efficiency. Harvard University Press.
Hani, A., Raed, S., Rawan, K. & Rand, A. (2019). Investigating the Impact
of Electronic Customer Relationship Management Success Factors on
Business Performance, 31(1), 105-127.
Jones, M. C. (2001). A Century of Stock Market Liquidity and Trading Costs.
Columbia University Mimeo, Retrieved from: http://pricing.free.fr/
docs/exec/jones.pdf.
Lyons, R. K. (2002). The Future of the Foreign Exchange Market.
Brookings-Whart on Papers on Financial Services.
Miikulainen, R., Liang, J., Meyerson, J. E., Rawal, A., Fink, D., Francon,
O., Raju, B., Shahrzad, H., Navruzyan, Duffy, N. & Hodjat, B. (2019).
Evolving Deep Neural Networks, Artificial Intelligence in the Age of
Neural Networks and Brain Computing, 293-312.
NYSE (2000). Shareownership 2000. Retrieved from:
http://www.nyse.com/pdfs/shareho.pdf.
Parisi, G. I., Kemker, R., Part, J. L., Kanan, C. & Wemter, S. (2019).
Continual Lifelong Learning with Neural Networks: A review, Neural
Networks, 113(8), 54-71.
Ray, T. & Charles, L. (1998). Forecasting S&P 500 Stock Index Futures
with a Hybrid AI System. Decision Support Systems., 23(1), 161-174.
Rime, D. (2003). New Electronic Trading Systems in Foreign Exchange
Markets. Norges Bank and Stockholm Institute for Financial Research,
470-504.
Artificial Intelligence (AI) in the Foreign Exchange Market 97

Stenfors, A. & Susai, M. (2018). High Frequency Trading, Liquidity


Withdrawal, and the Breakdown of Conventions in Foreign Exchange
Markets, Journal of Economic Issues, 52(2), 387-395.
Tygier, C. (1986). The Foreign Exchange Market: A Descriptive Study,
Harvard University, Retrieved from: http://www.pirp.harvard.edu/
pubs_pdf/tygier/tygier-p86-8.pdf.
van Liebergen, B. (2017). Machine learning: A revolution in risk
management and compliance? Journal of Financial Transformation, 45,
60-67.
Yoon, Y., Tor, G. & Swales, G. (1994). Integration Artificial Neural
Networks with Rule-Based Expert System, Decision Support System,
11(2), 497-507.
Yu, L., Lai, K. K. & Wang, S. (2005). Designing a Hybrid AI System as a
Forex Trading Decision Support Tool. The 17th IEEE International
Conference on Tools with Artificial Intelligence.

View publication stats

You might also like