Cash forecasting is a process to identify receivables and payables, which enables an enterprise to manage its liquidity position. Identification of receivables and payables is important because it allows the enterprise to manage its working capital cost by reducing bank overdrafts or lending and to optimise its use of surplus funds.
Being able to identify its obligations in advance means that an enterprise can optimise its cost of funding and investment opportunities from both a short-term and a long-term perspective.
Robust cash forecasting can lead to overall improvements in:
Various tools can be used for cash-forecasting purposes, including spreadsheets, specialist software and enterprise resource planning (ERP) systems. Alongside the tools that are used for forecasting, the strategy that is applied will also ultimately determine the usefulness of the forecasted information to the enterprise.
In most cases, it is not the cash-flow information such as currency, amount and account that are important; nowadays, an enterprise can derive much greater benefit from the transaction-, counterparty- and settlement-related information that can be associated with a cash flow.
The three main approaches to cash forecasting are:
1. Transaction life cycle-based approach
This is a strategic, enterprise-level approach that requires various business units, such as procurement, sales, HR, finance and operations, to be integrated via ERP or specialist software.
A single transaction reference number is allocated to the transaction and the transaction carries this number throughout its life cycle, during the course of various different processes.
For example, in the procurement cycle, there will be the following processes: raising of a purchase order, invoice generation, goods acceptance, invoice payment, partial payment return, vendor refund and so on. Some of these events would have a cash impact and some would not have any cash impact. However, the same transaction number would be carried over.
Each process relating to the transaction is expected to enrich the cash flow with updated relevant information that would be of interest to treasury. The beauty of this approach is that treasury could have complete cash, transaction, settlement and counterparty-related information associated with the life cycle using a single transaction number from the start until the end of transaction.
This is a huge benefit, since a single transaction number would unite multiple applications for various reconciliation purposes from transaction tracking through to instantly advising treasury of any obligation that the organisation has to consider, thereby providing enough lead time to respond.
Various tools can be used for cash-forecasting purposes, including spreadsheets, specialist software and enterprise resource planning systems
This approach is best suited for large enterprises with well-defined processes within business units that may exist in silos across multiple geographic locations. Implementation of this approach would lead to transparency on transactions, the purpose behind a transaction and the various status changes associated with it, counterparty information, the impact on foreign currency, any applicable netting opportunities, and receivables management.
With this approach, the enterprise has the freedom to define the life cycle based on the process maturity of the organisation and how it wants to use the details in the forecasted information.
For example, cheque processing can be segregated or combined with payroll processing based upon the organisation’s requirements. So, if an organisation wants a view on the cheques paid to employees and vendors, then it can track this by combining the two activities together. On the other hand, if it does not want to track the payroll cheque processing, then only vendor cheque processing would be implemented.
There may be cases where transaction-level forecasting may not be possible – for example, in the US, securities settlement takes place on a net basis, thereby individual-level transaction status monitoring may not be possible from an end-to-end perspective. In such cases, this model needs to be modified.
2. Event-based approach
An event-based approach is best suited to a regional business unit of a global enterprise or a medium-level enterprise. This is because it looks at the cash part of the transaction life cycle only.
Funding requests would be triggered, based upon a limited set of events identified in a transaction process, leading to the creation of cash advice. This cash advice would be sent with transaction-related information to the treasury of the regional business unit or medium-level enterprise.
A unique transaction number would be allocated that would link the transaction to the event that triggered it until the end of its life cycle.
The information carried by the cash flow is severely restricted by the availability of it at the execution of the event, since the transaction may not be carrying all of the vital counterparty- or settlement-related information at that point.
For example, in the case that a securities corporate action has been initiated in the form of dividends, then appropriate cash flow would be generated, since the dividends would impact the account balance. On the other hand, once the cash-flow event has been triggered, it would only contain information relating to the security and the client account details.
Counterparty information may not be available until the money is credited to the client account and, depending on the implementation approach, status tracking and reporting across application may not be possible. Also, treasury would only be aware of the cash-related impact once the event has taken place, reducing the lead time available in the life cycle-based model.
3. Impact-based approach
An impact-based approach takes care of the specific need of a business unit within an enterprise or a region.
Events would be identified and triggers created for advising treasury of the generation of cash funding. Then cash-flow information would be extracted based upon the defined events.
The information captured using this model is limited since transaction origination information may have been lost during the transaction’s life cycle. Also, the information is extracted from the transaction at the very end of its life cycle, which provides very little lead time for treasury to plan an adequate response in the most cost-effective manner.
This model is most appropriate for enterprises where processes are not mature and where cash forecasting is just required for a section of bank accounts or business units. In Asia-Pacific, this model is quite commonly used due to the more limited consolidation and size of enterprises that exist.
The three models cater to organisations at different levels of maturity and according to how well they would be able to use the information captured in funding requests. These models assume that funding would be requested at transaction level, however, there may be cases when funding requests are being extracted on a netted basis, which means they would lack transaction-related details and could not be further used for reporting and status update purposes.
Ajay Aditya has acted as a treasury consultant to global banks for more than 15 years and currently resides in Singapore
