Current systems inadequate, company forecasters claim... eight questions you can ask to help evaluate any system.

Cash flow forecasting systems presented in the existing literature fail to address a number of key issues that have a serious impact on such forecasts. Based on the experiences of Capital Cities, Communications, owners and operators of standard broadcast and cable television stations, trade publications, and general circulation newspapers in various cities, we have isolated eight major areas of concern.

The problem at Capital Cities is, perhaps, easier than at most other companies, because of the nature of some of its businesses, and because a number of its operations are regulated by Government agencies:

The Federal Government limits the number of stations owned by one company; local governments regulate cable television systems; advertising revenues of newspapers are affected by seasonal and event-motivated advertising such as elections; trade publications (Women's Wear Daily is an example) are affected by seasonal advertising.

Despite the relative ease of forecasting cash flow at Capital Cities, actual experience in trying to adopt systems offered by many sources demonstrated their inadequacies. It is from that experience that we drew up a list of eight questions forecasters should ask to help them evaluate their own system and those offered by outside sources.
Here are the questions along with some comments about how we answered them at Capital Cities:

1. Does the system formalize the costs and the benefits of cash flow forecasts?

The authors of some cash flow forecasting systems are unaware of the reasons anyone bothers to forecast. Such systems lack the necessary components to yield conservative forecasts. The rewards of being surprised in the future with too much cash usually do not equal the penalties for finding yourself with too little cash. The interest rates for short-term borrowings exceed the rates available on short-term investments. The Internal Revenue Service imposes penalties for the underpayment of estimated taxes by corporations (see Form 2220).

The amount of time and money a cash flow forecaster spends on refining a system depends upon the expected gains from increased expenditures. By simulating cash flow forecasts under different assumptions, one can determine which one or more of the outcomes are satisfactory.

2. Does the system permit you to prepare forecasts frequently?

Some proposed cash flow forecasting systems use annual data only. Managers of large firms cannot wait for one year to see the next forecast. Any system must permit the forecaster to prepare forecasts quarterly, monthly, or even daily if the nature of the business warrants rapid revisions in forecasts.

Many firms share with Capital Cities Communications the need to prepare monthly forecasts for these reasons:

1. Sales revenues are seasonal. For Capital Cities Communications, 4th quarter sales are highest and 1st quarter sales are lowest.
2. Debt obligations mature at different times throughout the year.
3. Interest payments on debt obligations are semi-annual for most firms, and even monthly for some.
4. Corporations pay income taxes quarterly in cash.
5. Investment revenues and interest costs of short-term borrowings can occur frequently throughout the year.

3. Does the system permit a comparison of aggregate and disaggregate forecasts?

Readers of annual reports of Capital Cities Communications can determine to some extent the difficulty of cash flow forecasting relative to other firms. By calculating financial ratios from the firm's aggregate data, users discovered that accounts and notes receivable (less allowance for doubtful accounts) were 12% to 13% of total assets in 1981 and 1982. By dividing net revenues by accounts and notes receivable, analysts determined that the average number of days receivables were outstanding for the same years ranged from 52 to 60 days (depending upon the choice of gross or of net receivables in the computation).

Surprisingly, some suggested cash flow forecasting systems use only aggregate data. Cash flow forecasters within a firm need to compare aggregate forecasts with disaggregate forecasts to evaluate which approach yields the better forecasts. For some firms, it is possible that forecasts with firm-level data are more accurate than forecasts based upon division-level data, because of the portfolio effect (too high and too low forecasts cancel).

For other firms, though, a forecaster may achieve more accurate forecasts and provide better explanations of forecast errors by summing division-level cash forecasts.

Unlike receivables, the forecasting of payables is an easy task for Capital Cities Communications. Payables are usually one-third of receivables and are paid in fewer than 60 days. Hence there is less need to refine the forecast by comparing aggregate and disaggregate forecasts.

4. Does the system rely upon operating budgets?

While some firms use operating budgets as motivational tools, operating budgets at Capital Cities Communications represent best forecasts. Cash forecasters have been able to use operating income as the starting point in the development of cash forecasts: Operating Income + Non-cash Items Operating Cash Flow + Non-operating Revenues = Total Cash Receipts - Total Cash Payments = Monthly Cash Excess (Deficiency) + Opening Cash Balance = Ending Cash Balance x Interest Rate = Interest Income (Expense).

If budgeted operating income is found to be unreliable, then it will be necessary to extend the cash flow forecasting model to forecast components of operating income.

5. Does the system use accrual or cash basis accounting data as inputs?

Non-cash items can be large dollar amounts for many firms. For Capital Cities Communications, some of these items are: Amortization of purchased goodwill, depreciation, and stock appreciation rights.
Cash flow forecasters use different definitions of cash flow. They disagree as to whether more accurate cash flow forecasts are possible using accrual or cash basis accounting data. Rather than assume that one approach is better, the better strategy would be to write computer programs to prepare two forecasts, one using one approach and the second the other.

Over time, the cash flow forecast can monitor the forecasts to determine which approach yields the more accurate forecast most frequently.

6. Does the system permit a variety of treatments for contingent items?

The directors of a firm may approve the acquisition of another firm under certain conditions. They may be willing to wait several years for the best opportunity to acquire another firm. Authors of proposed cash flow forecasting systems tend to ignore this major forecasting problem.

If the forecaster assumes the acquisition will never be possible, decision makers will still want to see a cash flow forecast under the assumption that the acquisition is possible. If a study of the firm's historical experience shows that, on average, there is a 3-year lag between approvals of acquisitions and actual acquisitions, the forecast could show cash needs 3 years hence or one-third of the total needs each year. Forecasting an acquisition requires a forecast of the acquired firm's operating income. Such presentations are difficult to explain and to support.

7. Does the system generate forecasts of Interest rates and of stock prices?

Accurate cash forecasts are especially useful if they incorporate accurate forecasts of interest rates and of stock prices. Decision makers rely upon good forecasting systems to make these decisions:

1. Issue bonds when rates seem unusually low, even if there is no immediate operating need for the cash.
2. Use excess cash to repurchase some of the firm's shares.
3. Invest in repurchase agreements or Treasury bills.
4. Roll over commercial paper holdings.

8. Does the system provide guidance for treating the effects of seasonal and cyclical variations?

Advertising revenues (expenditures) are affected by seasonal and cyclical patterns. At Capital Cities Communications, for example, these revenues are affected by national elections and the Olympics. Both change revenue patterns.

The economics of national and of local advertising revenues are very different. National advertising revenues may be strong during the same period local advertising revenues are weak, and vice versa. The characteristics of firms advertising at the national level are different from those advertising at the local level.

Using signed contracts as a means of forecasting advertising revenues poses a problem, because of the way such contracts are written. Advertising rates are based on the frequency of insertion in print media and on frequency of broadcast in radio and television.

Contract rates are quoted to advertisers on those bases. The rate actually charged, however, is determined at the end of the contract period based upon the actual frequency and space (time) used. This arrangement not only makes the amount to be collected uncertain but also delays payment.

Another problem stems from the tendency of advertisers to cancel insertions or to add them during the contract period. A partial solution to this problem is to compare the percentage change between any given point in time in the current year and the same point in time in the previous year. Percentage changes in budgets for advertising revenues provide a benchmark for tracking moves in actual amounts of orders on the books.

Information Sources
Cash management executives have at their disposal a number of sources from which to obtain guidance in improving cash flow forecasting techniques. Besides newsletters from Certified Public Accounting firms, there are numerous other publications such as: Cash, Cash Management Forum, Cash Flow Magazine, The Cash Management Newsletter, The Cash Manager, Journal of Cash Management, The Corporate EFT Report, The Leahy Newsletter, and Pension and Investment Age.
Also a recent book by William Loscalzo, Cash Flow Forecasting, reviewed in the Winter 1983-'84 issue of The Journal of Business Forecasting.

Reprinted from
The Journal of Business Forecasting, Fall 1984
Summary Information