- Jochen Schwabe
- General Manager and Partner, Schwabe, Ley & Greiner
Particularly during turbulent times, good risk management requires professional work rather than better forecasts or superior tools. In this article Jochen Schwabe comments on recent project-related experience.
The level of interest in the issue of FX management correlates with exchange rate volatility. Since the end of 2014/beginning of 2015, this issue has again been en vogue due to the extreme fluctuations of the US dollar, Swiss franc and the rouble, not to mention the recent devaluation of the renminbi. These events trigger the appearance of numerous specialist papers from so-called ‘experts’ who explain how the future policy of the Fed or the Chinese government should be interpreted or who advise us to rely on new, better and superior tools or strategies: more options, ideally structured, overlay strategies and the like. Two key messages here:
We have carried out more consultancy projects related to FX risk management in the last six months than in the preceding two years. This allows us to draw several conclusions.
1. Accurate risk identification and analysis are key
Every business model has its own pricing mechanisms. While oil companies pass on the volatilities related to the US dollar and oil prices to us consumers via digital price displays at filling stations in real time (and therefore are not themselves exposed to FX risk), car manufacturers bear risk for periods considerably longer than a forecasting horizon of one year, given that they can’t afford to correspondingly adjust the prices of their products in foreign markets annually.
If you don’t exactly know the mechanisms involved and, as a result, the actual exposures and risk profiles of your own company, you don’t have a basis for a reliable strategy. Our projects have highlighted the following: the risk profiles of many corporates are more complex than you might think. They are often also differentiated in terms of various business lines, products and markets. And what’s more: they are constantly subject to change, e.g., due to the influence of the internet, the opening up of previously protected markets, the market entries of new competitors, etc. While the standard risk horizon of a tour operator, for example, used to be the period covered by a single brochure, i.e., around six months, this no longer applies to at least that part of the business which is sold via the internet where it is possible and/or necessary, due to competition, to pass on changes in cost factors much faster.
Risk profiles tend in reality not to overlap with forecasting periods or financial accounting deadlines. Risk exposures also tend to arise not at the time that they are recorded but some time earlier, when a binding offer is formulated, a contract concluded or when prices are set. Prices are defined and change to reflect markets. It is therefore not possible simply to define risk horizons via policy documents issued from the relative isolation of the corporate treasury department. They need to be analysed ‘closer’ to the relevant transactions.
This might be tedious (see Case Study 1 for an example of the procedure) but is nonetheless essential. Without such an analysis or as a consequence of inaccurate results, any risk management efforts will not only be off-target but perhaps also contra-productive or even risk-exacerbating. If a business division with a price-adjustment horizon of three months in its markets is forced by a group-level policy to ‘hedge’ its forecast cash flows for the next 12 months, then the actual risk exposure is tripled and the relevant prefix reversed due to the fact that Months 4 to 12 were previously not exposed to any risk. For competitive reasons, the business division is forced to take exchange rate changes into account in its prices from Month 4 onward at the latest, despite having ‘fixed’ the originally forecast amounts, i.e., there exists a clear risk mismatch.
Case Study 1 – Risk analysis
The corporate treasury department of a major conglomerate with global operations wanted to review the assumptions about the risk horizons defined in the group-level FX risk policy document. This document stated that at least 80% of all forecast cash flows in FX were to be hedged for a period of 12 months. After a number of changes in terms of the group structure and the business model, doubts had arisen as to whether the uniform assumption of a 12-month risk horizon could apply to all business divisions in practice. It appeared challenging to fundamentally analyse this for all business divisions. We proposed initially analysing one business division where the doubts about the validity of this approach were particularly acute rather than analysing all of the divisions. Applying Popper’s falsifiability principle, the idea was that if the sample division highlighted that the ‘12-month policy’ was incorrect, then it would be possible to investigate other divisions step by step, and this would have also served as the justification for a more comprehensive project.
The first step was a workshop with personnel from the sales, purchasing and controlling departments of the relevant division, i.e., well away from the classic domain of treasury and at the heart of the underlying core business processes. After a two-hour meeting, it was already clear that such a long risk horizon was not appropriate in this case. It was obvious that prices changed in response to numerous factors, of which exchange rates were only one, over short periods ranging between six and eight weeks. It was plausible that this was associated with the nature of the business (oligopolistic situation) as well as the geographical locations of the production sites of relevant competitors (all in the same currency region).
Given that a copy of the minutes of the meeting, full of anecdotal evidence, did not appear sufficient in order to question the validity of the entire corporate policy and to trigger change, the actual sales prices of recent years were subjected to a multiple regression analysis. This procedure is necessary due to the fact that clear 1:1 relationships between exchange rates and product prices are rare. A whole series of factors generally influence prices, as a result of which, in simple terms, it is not possible just to claim: “A change in the exchange rate of x leads to a change of y in the price at time point z”. In reality: “A change in the price at time point b is, inter alia, also a result of the change in the exchange rate at time point a.” The most challenging aspects of such a procedure are often establishing constant and valid time series of the prices themselves as well as the influencing factors to be investigated. This was also the case here. The calculations took into account not only the relevant exchange rates and commodity prices but also the share prices of key competitors as approximations of their relative commercial strength.
The regression analysis clearly highlighted that a change in the exchange rate investigated had a highly statistically significant influence of more than 30% on the actual change in prices as early as one month later. As a result, the assumption of a 12-month risk horizon should be abandoned. Entering into FX futures for Months 2 to 12 didn’t hedge the risks that this division was exposed to. On the contrary in fact. This was an important lesson for the company in question and one which was acquired within just a few weeks and at the cost of just a few consulting days.
2. Accurate risk measurement is the basis of an efficient strategy
Once the risk-bearing positions have been identified, it is then possible to accurately calculate the risk exposures arising out of these positions today using any laptop. It is surprising that many companies still rely on ‘simplification’ (in other words, inaccurate) processes despite the fact that the right methods, ranging from variance and co-variance to Monte Carlo simulations, are now on hand at every workplace using readily available tools.
It is inaccurate, for example, to rely on popular sensitivity analyses which mostly consist of nothing other than the question of what happens if the USD/EUR exchange rate rises or falls by 10%. This ignores not only that a change of 10% today may have become more probable than it was yesterday, or the reverse, but also the fact that risk factors are always also correlated with each other. It is possible that the change associated with the US dollar is compensated by another change related to the renminbi. Such correlations within a portfolio need to be explained and measured in order to draw the right conclusions. It might sound practical to ignore correlations and, for example, to simply calculate the individual VaRs for the relevant currencies, but this is not objective and ultimately provides an inaccurate overview of the entire portfolio. Correlations exist, change over time and cannot be simply excluded and ignored.
This simplified approach often leads to an overvaluation of the actual risks by ignoring diversification within the portfolio. This claim is usually countered by the argument that it is better to overestimate risks in order to be prudent and conservative. But can this really be a satisfactory or professional approach for a financial officer? From the corporate point of view, however, is it really practical and desirable for all risk assessments to be overinflated and, as a result, to provide more risk coverage in the form of hedges, insurance or even equity than is actually necessary?
For FX risk management, accurate risk measurement, taking into account correlations between the various risk-bearing positions, is definitely important given that this is the only way of developing efficient strategies for the entire risk portfolio.
3. Efficient risk strategies are only possible at the portfolio level
We repeatedly make two observations in the course of risk strategy projects:
Firstly, apparently ‘practical’ approaches (“Let’s just hedge our three main currencies.”) are often inefficient and sometimes even risk-exacerbating when viewed more closely. The cause of this is of course normally to be found in the risk measurement methods described above. This also applies to other apparently practical approaches such as increasing the hedging ratio across the board based on the motto: “The more the better”. An ostensibly prudent increase in the hedging ratio from 80% to 90% in many cases delivers no marginal benefits and in some cases can even exacerbate rather than mitigate risks.
Secondly, often, too many individual transactions are performed. In such cases, thousands of micro-level hedges are concluded, with a corresponding volume of associated work, despite the fact that the same or better hedging could have been achieved with a handful of macro-level transactions. This finding is also dependent on the applying the right calculation methods (see Case Study 2).
Case Study 2 – Risk strategy
The CFO of a medium-sized trading company with global operations commissioned us to review the efficiency of the existing FX risk management practices. The currency positions and cash flows had already been very well organised and prepared internally. FX-related revenues in 22 different currencies accounted for around 50% of total turnover, with by far the largest single share of this being denominated in US dollars. Until this point in time, the company had pursued a relatively strict hedging policy covering all currencies, as a result of which even small amounts were individually hedged, prolonged, etc. In total, a five-digit number of individual hedging transactions were concluded every year.
After conducting a risk analysis such as that described in Case Study 1, it was clear that the risk horizon was three months, which immediately put the significance of the first observation (50% of revenues denominated in FX) into perspective and reduced the volume of the underlying positions.
These positions were captured using the overall risk tool developed by SLG (a simple tool based on Excel and a MatLab core for risk measurement at the portfolio level). Simulated across several time slices, it was revealed that, with a 95% degree of probability, the entire FX risk exposure associated with the portfolio amounted to just 2% of EBITDA. This appeared to be negligible to the board of management.
In order to back up this finding, a stress test was performed based on the assumption of the largest annual change in the exchange rate in the last 15 years for each of the 22 currencies, with these all occurring in a single period. Based on this scenario, the resulting risk was still only 7% of EBITDA.
In a final step, strategy simulations highlighted that the complete hedging of by far the largest single position in US dollars would have had absolutely no risk-mitigating effect due to this ignoring of the diversification effects. Halving the overall risk would have been possible by means of just a few co-ordinated hedging transactions in just a handful of currencies. Due to the low level of risk involved, the board of management actually decided to refrain from engaging in further hedges and to repeat this analysis at regular intervals. This will avoid future hedging costs and the credit lines of the company will no longer be burdened by hedging transactions, which was also one of the desired outcomes in this case.
Overall, it is clear that the professionalism of corporates in dealing with FX risk exposures has improved considerably over recent years. Meaningful risk KPIs are now already widespread, limits have been defined and systems for reporting are now available. In order to achieve further improvements, however, our recent project-based advice is the following: Don’t look beyond the bounds of your company for the answer in the form of better market forecasts or tools. The key to better FX risk management lies within; in the risk analysis of your business model and the use of the right methods.
