One of the perennial challenges facing treasury is to have a comprehensive overview of exposures at a level granular enough that they can be hedged in compliance with hedge accounting requirements. That means identifying what the exposure drivers are, when they’ll occur, how large they are likely to be, and how they’ll change over time. At a NeuGroup FX Managers’ Peer Group meeting late last year, members discussed these challenges, specifically as they related to ownership and gathering all the relevant data from scattered systems. Also, can’t robots fix it?
As for exposure forecasting ownership, who’s in charge? FX risk management is a highly centralized function, while the exposures are not, spreading out all over the world for large global corporates like the members of the FXMPG. This raises the question of who should own and be accountable for exposure forecasting.
One member at the meeting noted the significant training efforts it takes to raise awareness of what revenue and expense data treasury needs from local finance resources and why it makes a difference for cash-flow hedge results. With frequent staff turnover, training must be ongoing, this member noted. Another member found that having a treasury person doing the cash-flow forecast for the larger entities around the world is more accurate than having local finance people owning it, due to factoring and other forecast-moving items treasury can monitor better.
At the same time, how can forecasters arrive at a number given how scattered the data is? Balance-sheet forecasts are highly reliant on ERP data and if the relevant data resides in too many places, hedging balance sheets becomes impossible. One member said that until they get all their billing into one system, balance-sheet hedging is off the table. Another problem is the existence of “phantom balances” resulting in unexpected net re-measurement; they stem from accounting errors and they mess up the forecast. The dilemma: Should you hedge them, because they can have a real impact, or not, because they’re not real?
So, what’s next in the continuing drive to enhance the process and accuracy of forecasting? Robotic process automation (RPA) and artificial intelligence (Ai) are on everyone’s mind. One member noted an internal initiative to identify how bots can be used in treasury and whether a bot “could figure out the forecast.” But when mapping out exactly how the bot would achieve it, it looked like every system in the entire company would require some change or other. For some this would be an impossible task given that oftentimes changing one system is a challenge. It also sounds like, given the state of some programs, automating an imperfect process rather than perfecting the process first so it can be automated later. In the final analysis, just building a bot for hedging doesn’t seem to be in the offing any time soon.