What is forecast in business administration



Statement about future events, especially future economic values variables (e.g. used as an economic forecast, situation analysis or population forecast), based on observations from the past and on theoretically and empirically well-founded, comprehensible procedures and theories. Forecasts are mainly based on variables that cannot or can hardly be shaped by the person making the forecast.

basis Every forecast is a general stability hypothesis, which states that certain basic structures have an unchanged effect in the past and in the future.

Different:technological prediction.


1. Direct / indirect forecast: A direct or autoregressive forecast is when the values ​​of an economic variable are forecast exclusively from values ​​of the same variables in the past. In the case of indirect prognosis, the causal relationship between different variables is built into the prognosis of one variable; In this case, however, direct forecasts must ultimately be used again.

2. Qualitative / quantitative forecast: A qualitative forecast only mentions the type and direction of the development of economic variables; a quantitative forecast also deals with the extent of this development.

3. Point / interval forecast: In the case of a point forecast, a special future value is sought for an economic variable, whereas an interval forecast requires a range within which the future value is located with a high degree of "certainty", usually defined as at least 90% probability. In the case of the latter, a confidence range can also be specified (forecast interval).

4. Conditional / unconditional forecast: In a certain sense, every prognosis is conditional, i.e. to be understood as an if-then statement; absolutely unconditional forecasts are not possible. However, one can proceed in such a way that forecasts for one and the same variable are alternatively made depending on certain detailed conditions and the user is left to assess whether these conditions are met, for example in the case of population forecasts under different conditions with regard to the development of the births.

5. Single forecast / forecast systems: A single forecast is based on a single economic variable. A forecasting system refers to a set of variables that are forecast in their mutual connection.

6. Various Maturities of forecasts: Short-term forecast (forecast period up to two years); medium-term forecast (up to five years); long-term forecast (up to ten years); secular prognosis (over several decades or centuries).

7. Development forecast (information forecast, trend forecast): The company has no noticeable influence on the variables to be forecast (e.g. market development of personal computers as a whole, changes in customer behavior or changes in the distribution system).

8. Effect prognosis (instrumental prognosis, decision prognosis): Forecasting the effects of measures taken by one's own company (e.g. on variables such as sales, turnover depending on certain marketing measures).

9. Indicator forecast: Indicators are used to forecast developments. Indicators can, but need not, be causally related to the variable to be forecast. Indicators can be divided into leading, co-indexing and lagging indicators. The number of building permits issued is a leading indicator of demand in the construction industry.


1. At short term Forecasts, especially in the operational area, direct forecasts are preferred, especially time series forecasts using moving averages or using exponential smoothing; at medium term Forecasts are used econometric methods to extrapolate the trend or, for example in the case of market forecasts, the forecast using growth functions (logistic function; Gompertz function). If present, too seasonal components (Time series components) the trend is forecast on the basis of historical values ​​that have been subjected to trend adjustment; for forecasts of the Future value the seasonal component is then added appropriately. Indirect forecasts are mostly done with the help of regression analysis and econometric models.

2. Basically differentiated will: a) Quantitative forecasting methods: Based on mathematical procedures (e.g. trend extrapolation, indicator forecast, exponential smoothing).

b) Qualitative forecasting methods: Based on experience, knowledge and instinct; used in the absence of quantitative data (e.g. Delphi technique, expert survey, scenario technique).

3. Forecasts are often made as direct forecasts based on ARMA models (ARMA (p, q) process).


1. Assessment of forecasts can be done first qualitative and in advance respectively. Criteria are the economic-theoretical foundation, the compatibility of individual forecasts within a system, the availability of qualified historical data.

2. In addition, the assessment is often made quantitatively and in retrospect through a suitable global identification of the forecast errors that have occurred (average of the absolute and relative forecast errors; correlation between forecast and actual value; Theil's coefficient of inequality). However, the forecast errors that have occurred should not only experience a measurement, but also a root cause analysis.

See also situation analysis, economic forecast.