10 rules of business forecasting

Monday, page 9, main

10 Rules of Business Forecasting

By Chang Se-moon

As we pass the midpoint of the year, all businesses, different units of government, sporting events, and nonprofit organizations may want to review the status of their budget and prepare for the rest of the year by making forecasts and appropriate adjustments. Even individual households need to make forecasts to keep their budget on a sound basis. Regardless of the nature of forecasts, there are some rules that all forecasters may want to pay attention to.

Forecasting is much more than a statistical exercise for two reasons. One is that statistical models require forecasts of variables that go into the model; whereas forecasting these independent variables can be just as risky as forecasting the dependent variable directly. The other is that experts in statistical forecasting tend to lack common understanding of the broader conditions that underlie statistical models. It is not unusual to find that upper level managers have a clearer picture of the future than lower level forecasters. Rule No. 1, therefore, is that you need to understand the broad economic environment that will influence your forecasting.

Even if forecasting is needed at certain times of the year, the work of forecasting is year-round. Monthly data need to be collected on time with watchful eyes on their revisions of preceding periods. Note that some websites post monthly data for only a month or two. Rule No. 2, therefore, is that you need to update your own data file every month, not just once or twice a year when you actually make forecasts.

Rule No. 3 is that if you make local or business forecasting that depends on national forecasts, be sure to check at least two or three different forecasts because they can vary widely. For global forecasting, you may occasionally check CESIfo, the IMF, and OECD forecasts that are all free. For forecasts of the U.S. economy, you may check the Office of Management and Budget (OMB), the Federal Reserve Board, and the Congressional Budget Office (CBO) forecasts. They are all also free.

Rule No. 4 is that you need to be quiet. Do not seek publicity by announcing your forecasts without the permission of your supervisor or your client, whichever the case may be. I personally have never released the content of my consulting reports without the permission of my client or until my client releases the information to the public.

Rule No. 5 is not to second guess how your forecasts may be used. Forecasts, especially those requested by politicians, may be used well beyond budgetary preparation. Forecasts of the economic impact of many short-term events such as festivals and sporting events are used more for promotional purposes than for budgetary preparation. In some cases, appearance is more important than accuracy.

Rule No. 6 is that if you are making forecasts as an independent consultant, you may want to know that the price elasticity of demand for your forecasts is less elastic when the work is done for public sector officials who often have a separate budget to pay for your study. Public sector officials care more about convenience and accuracy. They are less sensitive to higher prices.

Rule No. 7 is that when you prepare forecasts, you might be thinking of publishing your methodology in an academic journal or a popular trade magazine. If you can do it once in a while, the benefits would be that your forecasts are likely to be more scientifically approached and that your forecasts will likely be more credible.

Rule No. 8 is that if you are in a position to gather information or data from another organization or department, it is important to be nice to everyone in the other organization, including the secretary or maintenance crew who appear to have nothing to do with the information. You never know.

Rule No. 9 is that all serious forecasters will have many sleepless nights because predicted values vary greatly with the assumptions that you make. If sleepless nights continue beyond a night or two, consider that time moves on and your forecasts become old before you start working on forecasts for the next period. No one cares about your forecasts as much as you do. Forget them and go back to sleep.

Read rule No. 10 below, only if you still feel lost and cannot figure out what to do next. Like so many good poems with anonymous authors, the author of the following also remains anonymous. If I recall correctly, the title of the poem is “How to Forecast Revenue”:

To find a value good and true,

Here are three things for you to do

Consider your replacement cost,

Determine value that is lost,

Analyze your sales to see,

What market value really should be

Now if these suggestions are not clear,

Copy the figure you used last year.

Chang Se-moon is the director of the Gulf Coast Center for Impact Studies. Write to him at: changsemoon@yahoo.com.