Grasping the Forecasting Nettle
by Robert Pehrson, Global Head of Liquidity Products
Cash flow forecasting has been at the top of treasurers’ lists of priorities for a number of years. For example, it was identified by 55% of treasurers in the SEB/gtnews Cash Management Survey 2007 as their primary objective. According to the same survey a year later, in the light of market events, forecasting was, for the first time, superseded by liquidity management. However, as the previous article highlighted, forecasting and liquidity are closely interlinked, and to be effective at gaining control over cash flow, and ensuring appropriate access to cash, requires treasurers to create regular cash flow forecasts. But why has forecasting been so elusive in the past, and how can this be resolved?
It is not difficult to see why forecasting has remained on the ‘to do’ list for so long. Despite everyone’s best intentions, there are always other priorities and opportunities that arise during the course of a year. For example, over the past ten years, treasurers have made substantial progress in centralising cash into both physical and notional cash pools, with demonstrable value to the business, as well as centralising financial functions such as payments. In contrast, forecasting has the disadvantage that the benefits are more difficult to measure, and tend to be progressive as the forecasting process is refined.
Forecasting is rarely a single process, and is handled in different places within the company. Even within treasury, short-term forecasts may be put together as part of the cash position process, while medium- to long-term forecasting is likely to be a separate process, often undertaken by different individuals. The problem of forecasting being a divided responsibility extends beyond treasury too. Many parts of the business will conduct cash forecasting for various purposes: sales, inventory management, etc. but there is rarely a common basis for constructing this information. This leads to a lack of consistency in presentation, the level of detail, and assumptions used, so those trying to construct and make decisions using the consolidated forecast are rarely confident in the trustworthiness of the data.
There are other challenges too relating to the fragmentation of the forecasting process (fig 1). Frequently, information is held in disparate systems, which store data in different ways, and it can be a complex process to integrate systems in a way that allows data to be transmitted reliably and consistently. Transferring data between systems could also lead to quality deterioration which again reduces trust in the consolidated forecast. Even when using a common systems infrastructure, such as an ERP, there are often different versions or instances in place, or the system does not cover the entire company, particularly in the case of companies with significant merger or acquisition activity.