A company can be profitable, well capitalised and still be caught short of cash at exactly the wrong moment. That is why corporate treasury is becoming less interested in whether a forecast exists and more interested in how reliably it predicts the timing and size of actual cash movements.
The shift is subtle. Forecast accuracy has long mattered to treasurers, but it is increasingly being treated as an operational risk signal rather than a purely planning-oriented KPI. The 2025 AFP Treasury Benchmarking Survey found that nearly three-quarters of treasury practitioners rank cash management and forecasting among their top priorities, while more than 60% identify cash or liquidity forecasting as their most challenging task. That combination—high importance and persistent difficulty—is exactly the environment in which forecast error becomes strategically informative.
The term “cash-flow forecast error” is not an IFRS-defined measure or a standard regulatory ratio. In this article it means the measurable difference between forecast cash flows and actual cash flows over a defined horizon, together with the bias, volatility and timing patterns behind that difference. The evidence supports the importance of forecasting; the inference is that leading treasury teams will increasingly formalise forecast error as a risk metric in its own right.
Why a forecasting miss is becoming more expensive
When financing was exceptionally cheap and liquidity buffers were plentiful, a forecast that missed by a few days could often be absorbed through excess cash, an undrawn revolver or short-term borrowing. The economics are less forgiving when borrowing costs are higher and capital is being scrutinised more closely.
The ECB’s second-quarter 2026 Survey on the Access to Finance of Enterprises shows that firms still use financing heavily for inventories and working capital: 40% of surveyed euro-area firms identified those purposes as their most common use of financing, while 37% cited fixed investment. The same survey found firms reporting further tightening in bank-loan pricing. A forecast miss that forces unexpected borrowing is therefore not just a spreadsheet error; it can alter financing cost and timing.
The ECB’s July 2026 Bank Lending Survey similarly found that corporate loan demand was supported by inventories and working capital, fixed investment among large firms, and debt refinancing or restructuring. In the United States, the Federal Reserve’s July 2026 Senior Loan Officer Opinion Survey reported stronger demand for commercial and industrial loans from large and middle-market firms. These are not direct measures of forecast error, but they show why unexpected liquidity needs can carry a meaningful price.
The risk is often timing, not solvency
Treasury forecasting is unusually sensitive to timing. A customer payment expected on Friday but received on Tuesday may have little effect on annual revenue and still create a large short-term cash gap. A tax payment can be known with near certainty in amount but not incorporated into the right entity-level forecast. Capital expenditure can be approved months in advance yet settle differently from the payment schedule assumed by treasury.
This explains why cash forecasting is different from an income-statement forecast. Under IAS 7 Statement of Cash Flows companies classify realised cash flows into operating, investing and financing activities. Treasury, by contrast, is trying to anticipate those flows before they occur and often at a much finer time resolution. A forecast can therefore be economically useful even when it is never published externally.
The distinction is receiving more attention at the standard-setting level too. The IASB’s Statement of Cash Flows and Related Matters project is examining improvements to the disaggregation and transparency of cash-flow information. That project does not prescribe internal cash forecasts, but it reinforces a broader point: understanding the composition and timing of cash matters, not merely the closing cash balance.
What forecast error can actually tell management
A single percentage-accuracy score is rarely enough. Treasury teams can learn more by decomposing error into several dimensions: whether forecasts consistently overstate or understate cash, whether error grows sharply at longer horizons, whether particular business units repeatedly miss, and whether misses come mainly from receivables, payables, tax, payroll, capex or financing flows.
Bias is especially important. A forecast that is “wrong on average” can still be dangerous if it is systematically optimistic. Persistent overforecasting of receipts can lead a company to invest excess cash that does not actually exist, repay debt too early or underestimate the size of required liquidity buffers. Persistent underforecasting has a different cost: idle balances, unnecessary borrowing capacity and weaker returns on cash.
This is why forecast error can become a risk metric rather than just a performance metric. A risk metric is useful because it changes a decision. If forecast volatility rises beyond an agreed tolerance, treasury can hold more liquidity, shorten investment maturities, delay discretionary outflows, increase committed facilities or investigate the operational source of the miss.
The data problem sits upstream of treasury
Many forecasting failures are not caused by the forecasting model itself. They begin with fragmented data. Sales teams know expected customer receipts, procurement teams know supplier timing, tax teams know statutory payments, HR knows payroll, and business units know capex milestones. Treasury is often the point at which these separate estimates have to become one cash view.
AFP’s cash-forecasting guidance explicitly highlights the need to assess both the accuracy and completeness of source data, including data from treasury systems, company planning documents, accounting, shared services and operating companies. It also notes that some data may not be available early enough or may not be reliable enough for the required forecast horizon.
That creates a useful diagnostic. If a business unit has consistently higher forecast error than its peers, the problem may be local receivables behaviour or poor submission discipline rather than the group model. If error spikes around month-end, the company may have an accounting-to-cash timing problem. If error is concentrated in one currency or legal entity, the issue may be structural cash visibility.
Automation can improve the process without eliminating uncertainty
Technology can reduce manual error and improve pattern recognition, but it cannot make uncertain cash flows certain. The most useful role of automation is therefore not to promise perfect forecasts; it is to produce better, faster probability estimates and to learn where forecasts tend to fail.
A practical example comes from ASML. An AFP case study on the semiconductor-equipment company describes how its treasury team worked with data-science colleagues on an automated AI model for foreign-exchange exposure forecasting after a process that had been labour-intensive and prone to error. The case is company-reported and specific to FX exposure rather than total cash forecasting, but it illustrates the direction of travel: treasury models are increasingly being built to measure and reduce repeatable forecast error.
The strongest systems are likely to combine deterministic information—such as scheduled debt service, payroll and tax dates—with behavioural estimates for more uncertain items such as collections, supplier timing and project payments. The goal is not a single perfect number. It is a distribution of plausible outcomes that treasury can manage.
From point forecast to liquidity range
This leads to an important counterargument: a company may become too focused on accuracy. Forecasts are always conditional on assumptions. A business with volatile demand, acquisition activity or commodity exposure may have a wider natural range of outcomes than a subscription business with recurring receipts. Penalising treasury for uncertainty outside its control can encourage false precision.
A better approach is to separate forecast quality from forecast certainty. Treasury can ask whether the model was well calibrated to the information available at the time, whether the error fell within the expected range, and whether the organisation responded appropriately when actual cash moved outside that range.
That is why scenario analysis matters. A central forecast may be 90% likely to cover ordinary operations, while a downside scenario shows that a 10% delay in customer receipts would require a revolver draw. The risk insight lies not in proving which outcome is “correct” today but in knowing what action is required under each one.
Banks and fintechs can sell the feedback loop
For banks, cash-management technology is increasingly a data business. Real-time balance feeds, virtual accounts, receivables matching and payment analytics can help clients compare expected and actual cash movements more quickly. That creates an opportunity to move beyond transaction execution and become part of the forecasting feedback loop.
Fintechs face a similar opportunity, but with a higher burden of proof. A forecasting dashboard is easy to demonstrate; proving that it reduces liquidity buffers, borrowing costs or forecast volatility is harder. The most credible tools will therefore expose model confidence, data lineage and forecast error over time rather than presenting a single opaque prediction.
For lenders, persistent forecast error can also become useful soft information. A borrower whose short-term liquidity forecasts are consistently unreliable may be operationally weaker than its headline leverage suggests. That does not make forecast error a credit score, but it can inform discussions about revolver sizing, covenant headroom and liquidity planning.
Investors should distinguish cash surprises from cash weakness
Investors often focus on quarterly free cash flow, but a cash-flow surprise can have several causes. Some are economic, such as deteriorating collections or inventory build. Others are timing effects that reverse quickly. A management team that can explain forecast accuracy and the drivers of variance may therefore provide more useful information than one that simply repeats an annual cash target.
The 2025 AFP Liquidity Survey showed that companies continued to prioritise safety and liquidity in short-term investment decisions. If treasury becomes more confident in the timing of cash needs, companies may be able to hold less idle liquidity or invest it for slightly longer without sacrificing resilience. The economic value of accuracy is therefore partly the option to use cash more efficiently.
Evidence versus inference
The evidence is strong that cash and liquidity forecasting is a top treasury priority, that practitioners still find it difficult, and that firms continue to use financing for working capital, investment and refinancing needs. The evidence also shows increasing use of automation and more attention to cash-flow transparency.
The inference is that “forecast error” itself will become a formal treasury risk metric across more companies. There is no universal standard requiring this, and many businesses already track forecast accuracy in different ways. The strategic argument is that measuring error systematically creates a bridge between forecasting, liquidity buffers and financing decisions that becomes more valuable as cash moves faster and capital carries a higher cost.
Conclusion: the forecast is becoming a control system
Corporate treasury is moving away from the idea that a cash forecast is simply a best estimate of the future. The more useful question is whether the organisation understands the range of possible outcomes, knows where its forecasts tend to fail and has enough liquidity to absorb those errors without destroying value.
That turns forecast error into something more than a retrospective score. It becomes a measure of how much uncertainty sits inside the company’s operating cash engine—and how much financial protection that uncertainty requires.
References
1. AFP — 2025 Treasury Benchmarking Survey Report
3. AFP — An In-House AI Solution to Improve FX Exposure Forecasting Accuracy
4. AFP — 2025 Liquidity Survey
5. ECB — Survey on the Access to Finance of Enterprises, Q2 2026
6. ECB — Euro Area Bank Lending Survey, Q2 2026
7. Federal Reserve — July 2026 Senior Loan Officer Opinion Survey
8. IFRS Foundation — IAS 7 Statement of Cash Flows
9. IFRS Foundation — Statement of Cash Flows and Related Matters
10. OECD — Global Debt Report 2026: Corporate Debt Market Outlook
