Redbridge invited three international treasury professionals based in the US, UK and France to share their views on successful cash flow forecasting. Today, Ferdinand Jahnel, VP, Treasurer at Marsh & McLennan Companies explains how Artificial Intelligence (AI) could improve the process of cash forecasting.
– Why does cash flow forecasting usually rank among corporate treasurers’ top priorities?
– Ferdinand Jahnel, Marsh & McLennan Companies: All of that is tied to liquidity management. Within a large organisation, corporate treasurers are responsible for the cash profile of the company. That means having a good understanding of the low points and high points in cash generation. It also means finding ways to bridge any gaps by issuing short-term debt or by relying on some type of committed bank line base. Moreover, when cash needs exceed the annual trends, treasurers need to have a broader understanding of how to manage the firm’s capital structure and what the long-term debt-to-equity mix should be.
– In what ways have you used treasury forecasting in your current and past roles?
– I worked for several companies before joining Marsh & McLennan, the world’s leading professional services firm in strategy, risk, and people. For instance, I worked for a healthcare distribution organisation, a software company and a manufacturing firm. I found across each of these industries that the best cash flow forecasts are generated from treasury, not from FP&A (financial planning and analysis) controllership or the people in the field. Treasurers have access to the best (historical) data on how cash flows typically trend within a company.
– How do you prepare your cash forecasting?
– There is a very idiosyncratic seasonality in every business. For example, at Marsh & McLennan, we know that we have a low cash point in the first quarter. During the rest of the year, cash rolls in almost like clockwork. Therefore, the way we ultimately develop our forecasts is by taking historical data on a daily basis and trending it with growth expectations. This methodology works well for us since renewals of annual insurance consulting contracts are fairly repetitive year after year.
– Which method of cash flow forecasting do you prefer?
– The indirect method is not going to help you with planning flows in treasury. It starts with net income and makes all sorts of adjustments, and that’s a somewhat abstract mechanism to reconcile a beginning and an ending cash balance on a balance sheet.
With the direct method, we can truly see which cash inflows relate to specific customer receipts or businesses and which relate to specific activities on the disbursement side, like payroll, tax, and so forth.
– Are you seeing any disruptive technologies that could improve the process of cash forecasting?
– We are trying to find applications for AI or machine learning where we hopefully see some ability to automate a process that I have described before as somewhat archaic. It is still manual. We still run a huge spreadsheet that we update and roll over into a new year quite manually.
Again, using that data makes a lot of sense for us because of the predictability of these cash flows. I would assume that especially repetitive patterns either on the cash in or outflow side could be something that machine learning or AI could do for us in renewing and updating these cash flows on a rolling basis going forward. I think these applications will eventually come to fruition, especially in businesses that have these somewhat plannable cash flows.