Accurate and timely information about what is likely to happen to the economy and society in the future has always been of value to business decision makers. One of the best-known stories of such forecasting is recorded in the first book of the Bible. In that case, Joseph was given the ability by God to interpret the Pharaoh’s dream and forecast that there would be seven years of very good harvests and then seven years of famine. Acting on that forecast, Egypt stored grain during the good years and survived the famine – and even prospered as people from surrounding lands had to come to buy food. More than a millennium later, the Oracles of Delphi also appealed to the gods to predict the future for the Greek kings and Roman emperors. In fact, there are reports throughout history of unusual forecasting techniques – often shrouded in mysticism.
While current economic forecasts may still seem to some to be mystically derived, today’s economic forecasters tend to rely more on data, computer models, and economic theories rather than divine inspiration although, given the accuracy of their forecasts compared to Joseph’s, one might question the change in tactics. However, forecasting has become an important part of planning for any other business. For example, sales forecasts impact the inventory of both finished goods and raw materials, the need for some types of personnel, space requirements, and financing, among other things. Macroeconomic forecasts may influence whether and where a business decides to expand, and the type and amount of financing that is used.
The development of modern economic theories such as the business cycle theory, the creation of various indexes, including the indexes of leading and lagging economic indicators, and sophisticated computer programs have given rise to new forecasting techniques. One of the key assumptions for most forecasters is that the past serves as the most important guide to the future. That does not mean that the future is a re-run of the past or that data about the past should be the only basis for a forecast. Obviously, socioeconomic conditions and global economies do not remain constant over time. Nevertheless, data about past trends and activities, when fit into a theoretical framework, provide some of the best information available.
Getting Started on Company Forecasting
If you are one of those for whom a business forecast usually means an intuitive guess, the following steps can help you begin a more formal forecasting process.
- Identify the problem.
- Determine how you would use the forecast to deal with the problem.
- Select the particular items you would need to forecast.
- Determine the appropriate time horizon for a forecast that deals with this problem.
- Research the techniques and theories used by others to forecast this variable in the past.
- Evaluate all possible opinions and consider pros and cons.
- Use a forecasting model that fits your business given your constraints and limitations.
- Make the forecast.
- Interpret the results.
- Make decisions and take action based on results.
- Implement a revolving recap of your forecast versus the actual figures.
- Modify your forecasting model or technique accordingly.
Forecasting Models and Techniques
To be able to follow the above steps, it would be helpful to have at least a brief introduction to the macroeconomic indices that are available and to the forecasting techniques that are commonly used by individual businesses. Beginning with the latter, there are two major ways to categorize forecasts: by the time frame to which the forecast applies, and by whether the techniques used are basically quantitative or qualitative.
The time frame for a forecast can be short-range, medium-range or long-range. Short-range forecasts typically cover the immediate future and are used to deal with issues of daily or weekly operations of a business. Typically, a short-range forecast would cover a period of one or two months. A medium-range forecast usually covers the period from one to two months to a year and is generally related to something like a yearly production plan. A long-range forecast would be for more than one or two years and is used to plan for the production for new products or the expansion of production capacity, or in the consideration of long-term financing.
Qualitative forecasting techniques are based on expert opinion and judgment. Today they are most frequently used when no quantitative data are available. When well done by a knowledgeable expert, qualitative techniques can provide reasonably good forecasts for the short term because of the familiarity of experts with the issues and problems involved. The primary problem in using qualitative methods is identifying the appropriate employees and then getting them to agree on a common forecast.
The main qualitative forecasting techniques are: executive committee, the Delphi method, surveys of the sales force, surveys of customers, historical analogy, and market research. Some of these obviously have a quantitative component as well. The executive committee technique consists of selecting a group of employees to represent the relevant departments of the firm and charging them with the task of creating a forecast, for example a sales forecast. They use inputs of various sorts from all parts of the organization, as well as from outside the organization, to create the forecast. This is probably the most common forecasting method used for both new and existing products and services. It is also one where strong personalities, or people in higher positions may dominate, and therefore the actual degree of consensus may be masked. In using this form of forecasting, it is important to be sure that all relevant information is solicited and considered.
The Delphi method is a variant of the executive committee approach, but the interaction is indirect, iterative, and very structured. It was designed to minimize the adverse effects of interacting groups. While there are variations of the method, the basic premise is that once the experts are identified, each is given a reasonably structured set of questions or issues to which to respond. Responses are sent to a coordinator or monitoring group that does not actually participate in the Delphi process. Responses are then fed back to each respondent, without identifying any individual with any opinion. Participants respond again in light of the views of their fellow respondents. There may be several repetitions of the process until a consensus emerges. The group may be brought together at the end to reach final consensus. Using electronic technology can help shorten the required time.
When conducting a survey of the sales force, data are gathered from members of sale force representing different products, segments or regions, and these data are combined. Managers may need to modify the aggregate numbers to ensure realistic estimates since there may be temptations to bias the reports. If rewards are based on beating estimates, the estimates may be very conservative. If the system rewards very high goals, the results may be very optimistic. To be effective, a good communication system should be available.
Direct contact with the organization’s customers is another way to make a sales forecast for a product or group of products. This method is preferred if the organization has relatively few customers. Mail questionnaires, focus groups, telephone interviews, or field interviews are tools used in market research. These inputs are used to analyze the market reaction in a target region, or outlet. These can be used for forecasting sales of a new product. Another method to estimate future sales of a new product is the use of historical analogy, where the forecast is based on the pattern of a similar product’s sales.
Quantitative Forecasting Models
These techniques use statistical methods for projecting from historical data. Some quantitative methods involve using company data to forecast individual firm performance, but there are also numerous indices and indexes published by the government or by private forecasting firms that can be of great help. In general, quantitative techniques are preferred when appropriate data are available. The main assumption is that the historical pattern will continue into the future.
Time series techniques are the most popular quantitative method. Two major types of time-series methods are moving average and exponential smoothing. The moving average is, as the name implies, a series of arithmetic averages. For example, to predict sales for the next period using a moving average, you would add up the actual sales for a specified number of periods (e.g., weeks or months) and then divide the total by the number of periods used. In a weighted moving average, weights are assigned to the previous periods. The sum of the weights must be equal to one, and more recent periods are weighted more heavily than those that precede them.
Exponential smoothing is a form of weighted moving average in which the new forecast is a weighted sum of the actual observed sales or other variable in the current period and the weighted forecast of that variable for the period. It therefore implicitly incorporates all of the data used to create the previous forecasts, but is much easier to compute. The major advantage of the moving average method is that it is easy to use, quick, and inexpensive. The major disadvantage is that it does not react well to variations that occur for reasons such as cycles and seasonal effects.
Indicators also are important in forecasting. Three types of indicators have been used for many years to help forecast the national economy. They are the Leading, Coincident, and Lagging indicator series. The terms refer to the nature of their relationship to the business cycle. The Bureau of Economic Analysis (BEA) of the Department of Commerce collects the data and has identified a particular series number for each indicator, 120 in all. As no single indicator series can encompass the economy’s diverse activity, it is necessary to track many different series to determine the direction of the economy.
Using these data series, the National Bureau of Economic Research (NBER) has developed composite indexes that capture and smooth the data contained in several of them including the three major Composite Indexes – Leading, Coincident, and Lagging. These indexes are reported by the BEA every month in the Survey of Current Business. Components of the Leading Indicators index include such things as average weekly hours worked and claims for unemployment insurance, manufacturer’s new orders, stock prices, orders for plant and equipment, an index of consumer expectations, the real M2 money supply, etc.
The Composite of Leading Indicators is useful for understanding the business cycle and is primarily intended to identify changes in the direction of the economy. Components of the Index of Coincident Indicator’s are: employees on nonagricultural payrolls, industrial production, personal income less transfer payments, manufacturing and trade sales.
The Lagging Indicators composite includes changes in labor costs per output, ratio of inventory to sales, and figures on installment credit and loans, among other items. In practical attempts to forecast the future, these Indices are among the most important tools available to most organizations, including the national government.
Any business operates on the basis of some forecast about the future, whether it is an intuitive guess or a sophisticated model involving large amounts of quantitative data. Perhaps in some small, stable businesses the former may work reasonably well but, the chances are, the more thoughtful and thorough the forecasting process, the more useful it will be. The techniques presented here are among those most frequently used today. For the most part, they are complementary. When carefully selected, and learned well enough to be properly applied, they can help a business move confidently into the future.