FX exposure management – a new automated approach to solve an old problem
The importance of FX risk management has grown significantly in recent years on the back of market volatility, increased geopolitical risk, and also regulatory and accounting changes. International expansion associated with geopolitical risk increases cash flow at risk due to FX exposures. This has meant that corporate treasurers face greater pressure to capture all FX exposures in a timely and accurate manner to achieve the best hedging strategy and results, thereby adding value for the corporate and at the same time reducing P&L and balance sheet impacts.
Many corporates are still finding it difficult to identify the FX exposures they face in a timely and accurate way. The exposures are normally submitted manually or captured by different business areas using unstructured and inaccurate methods. Most of them are only submitted when a purchase order or invoice is raised where risk is already crystalized and there is increasing cash flow at risk. This also results in treasury departments being reactive instead of proactive strategic partners to the business. The key underlying reasons for this are not only technology related – far from it. Rather, we see the following main hurdles in FX exposure management:
- a silo-based approach;
- lack of accountability from other units and departments;
- different processes, templates and procedures outside treasury responsibility; and
- a scattered IT landscape.
Some corporate treasurers say that most of the items stated above are outside treasury ownership and control. This is true: a treasurer cannot simply change an accounts payable, procurement or FP&A process, nor can they change the enterprise resource planning (ERP) landscape. But recently things have changed with new technology in the market. So instead of waiting for others to change, why not take ownership of the FX exposure process accuracy and automation?
To collect short-term crystalized exposures, one option is to build interfaces between several systems (ERP, budgeting and planning systems) and the treasury management system. Another is to ensure that all businesses and key stakeholders submit their short- to long-term forecasts in a timely and accurate manner. Both methodologies still continue to thrive in the corporate market, but both have limitations:
- Interfaces are costly, require high maintenance and monitoring, and are suitable for a corporate with a static business or IT landscape.
- Manual submission of exposures is error prone and costly. The forecasting processes normally tend to require the involvement of a lot of FTEs for gathering data, forecasting, validation, control and variance analysis.
Areas to explore
But how can a corporate treasury increase accuracy, quality and timeliness of their FX exposures without building interfaces or being dependent on business units? From an exposure point of view, there are currently several areas a corporate treasurer can explore that are more efficient and accurate than interfaces or manual submissions, and far less costly. These include:
- Robotic process automation (RPA) – For a short-term FX exposure forecast, RPA is less costly and more flexible that building interfaces. It can be used to gather short-term exposures from different ERPs and source systems. It can also perform other required activities in the FX exposure management process such as validation of thresholds and variance controls, consolidating information into one source, performing calculations, getting market data (FX rates, forward points) and getting budgeting and planning data.
- Predictive analytics – for a short- to long-term FX exposure management, using predictive analytics will increase forecast accuracy and reduce FTE costs. Predictive analytics can be used to trend cash flows and FX exposures based on existing historical data, forecast data but also other criteria and assumptions, when required. It is already used extensively in financial institutions to estimate prepayment cash flow or probability and early warning of loan defaults, and can also be used within corporate treasuries in a similar way. For implementation, different methodologies can be applied, such as: embedded in the system; bespoke development; Excel based. In addition to this, different types of machine learning techniques can be used:
- Supervised machine learning (ML) algorithms can be used to establish an automated baseline business volume forecast as well as perform a risk assessment on it (cash flow & earnings at risk), most common for corporates.
- Unsupervised machine learning techniques can be used in support of that goal, e.g. to identify correlations between business, financial and economic measures as well as detect anomalies in a business units’ manual data entries.
Approaching a complex project
From a corporate perspective, the main benefits of achieving global visibility of short- to long-term cash flow forecast and FX exposures are clear. Two of the main objectives (improving accuracy and reducing hedging costs) are easily measurable, with direct benefits linked to those. Luckily, new technology can assist in achieving those objectives.