AI in corporate foreign exchange hedging
March 2018 | SPECIAL REPORT: MANAGING RISK
Financier Worldwide Magazine
March 2018 Issue
Foreign exchange (FX) hedging is a complex and time-consuming endeavour for corporates. To come up with an optimal hedging decision, numerous parameters need to be accounted for. Aside from the market environment, a company’s business model and activities play a vital role. In the volatile economic reality of today’s business, trying to keep track of all relevant parameters to consider in FX hedging can be an impossible task.
Day-to-day business activities of an internationally operating corporation lead to FX exposure. Receiving or making payments in any currency other than the balance sheet currency of the company creates risk and thus requires some sort of hedging. In order to reduce risk and to create planning security, most companies implement static corporate hedging policies that do not account for shifting markets or corporate environments.
Additionally, the business and market environment determining optimal hedging strategies may correlate. However, most businesses consider the market environment only, disregarding the greater picture by ignoring any correlation to the business environment.
Conventional FX hedging
So how do corporates handle these risks? The answer is: very differently. Cash flow hedging is mainly driven by internal guidelines rather than current market conditions. Additionally, these hedge policies leave plenty of room for interpretation complemented by human decisions. Market components that distinguish basic from sophisticated hedging strategies are complex and costly to keep track of. Many of these resulting strategies are based on heuristics and gut feeling.
More sophisticated hedging strategies require know-how which diverts from the core competencies of the corporation and are costly without necessarily generating sufficient additional value. However, even the most complex hedging strategies leave room for improvement. The reason for this is the inability of the human brain to comprehend and process all relevant information that concerns not only internal corporate parameters but also those of the market environment.
Using AI to hedge FX risk
The implementation of an optimal hedging strategy is a seemingly impossible task due to the numerous influencing parameters and the limited information processing capacity of human decisionmakers. In addition, against the background of a dynamic corporate environment and volatile financial markets, a constant reassessment of the previous decisions must be made. Luckily, recent technological progress allows us to tackle these limitations: the increased availability of market information, the use of Big Data technology and digitised processes can help state-of-the-art artificial intelligence (AI) to support corporates in making optimal hedging decisions.
In contrast to the traditional FX hedging activities of companies, with the use of AI, the decision is no longer left to a subjective human decisionmaker. Instead, a complex algorithm can be used as a ‘robo-adviser’ for FX hedging. In order to reflect the economic situation in the best possible way, the algorithm can be directly linked to real-time data. Based on the synthesis of market data as well as customer specific information, the user is given a proposition on the amount and timing of the hedge. The algorithm optimises the hedging decisions in such a manner that market opportunities are used in the individually optimal way while the internal conditions of the company, e.g., the risk limit and minimum hedging ratios, are always strictly adhered to.
To ensure that the economic reality is sufficiently well represented, a large amount of scenarios needs to be simulated. No matter how likely or unlikely the occurrence of a particular scenario, the robo-adviser will always find the optimal strategy. In addition, the robo-adviser can be seen as a self-learning system, using only information that is actually relevant to finding the optimal hedging strategy.
Therefore, the AI ses ex-post evaluation of the previous strategy with the help of a reward function. Based on this function, the algorithm constantly adapts and improves itself. Here, not only the current market situation is taken into account, but also a path dependency resulting from the historical development of relevant parameters.
From a treasurer’s point of view, one of the main advantages of using AI is the ability to include the economic reality of the company in the decision making process. For example, consider a company with the euro as its balance sheet currency, exporting goods to the US creating FX exposure. Let us assume that the economic situation of the company develops positively, leading to higher sales, thus increasing the US dollar exposure. This may generate opportunities on the FX markets since the risk-bearing capacity has also improved. Taking these circumstances into account, an AI-based hedging approach may recommend a lower hedge ratio compared to a hedge ratio in a strained economic situation. Although this approach seems intuitively plausible, it is common practice in conventional FX hedging to use more rigid policies which completely disregard the impact of the company’s economic reality.
Unlike corporates, institutional investors such as asset managers or investment funds face a very different kind of FX exposure. While corporations are mainly concerned with cash flow hedging, institutions face the risk of FX volatility nullifying their offshore investments’ performance. Since fund value may be measured in another currency a decline in the investment currency may swallow any positive performance the investments may have had or even worse decrease the fund’s value altogether. Ergo, hedging policies and requirements are fundamentally different but nonetheless just as vital. Therefore, AI-based technology could also create value in taking the correlation of the underlying assets and FX markets into account.
With AI becoming more vital in the world of finance, corporates face a new solution to addressing FX risk. Not only does AI lead to less subjective decision making, it also provides the opportunity to continuously monitor and adjust previous hedging strategies. With AI, corporates have a powerful tool at their disposal allowing them to address FX risk in an efficient and individually optimal way and to divert time and effort to the company’s core competencies.
Roland Probst is a managing partner and Christoph Johanning and Philipp Faulstich are consultants at Lucht Probst Associates. Mr Probst can be contacted on +49 (0) 69 9714850 or by email: firstname.lastname@example.org. Mr Johanning can be contacted on +49 (0) 69 9714850 or by email: email@example.com. Mr Faulstich can be contacted on +49 (0) 69 9714850 or by email: firstname.lastname@example.org.
© Financier Worldwide
Roland Probst, Christoph Johanning and Philipp Faulstich
Lucht Probst Associates