# Forecast (statistics)

The forecast in statistics is the estimation of what is expected to happen with respect to a variable. This, based on a numerical analysis.

In other words, the forecast is a way of anticipating what may happen in the future, always using mathematical tools.

The forecast can be used at different levels, allowing decisions to be taken to mitigate the negative effects that certain contingencies may have if they materialize.

For example, a company can forecast expected sales for the next year, considering a negative scenario, a moderate one and an optimistic one. For each of them, suppose that the estimated variation in income is 1%, 4% and 10%, respectively. To arrive at this result, many variables, both internal and external to the organization, and, of course, historical data were included.

Similarly, a country's monetary authority can make a forecast of how much gross domestic product (GDP) could grow during the second quarter of the year. This, compared to the same period of the previous year. This forecast should be based on a comprehensive econometric study.

With the above, we can infer that a forecast is objective and not intuitive, as a prediction could be.

That is, the prediction could be based on logic or a theoretical framework, but the forecast is based on a numerical analysis.

## Forecast types

Forecasts in a company or project can be classified as follows:

- Short-term forecasts: It is valid for one year. It is often used to plan the supply of inputs, calculate the level of production and for the allocation of labor to different tasks.
- Medium-term forecasts: They can have a term of six months to three years. They are used to estimate sales, production, and cash flow. Likewise, they allow structuring budgets based on medium and long-term objectives.
- Long-term forecasts: They have a horizon of more than three years. They are often used to estimate the results of certain investments, study the launch of new products, evaluate market trends and, in general, develop long-range projects.

## Forecasting tools

Some tools for forecasting are as follows:

- Confidence intervals.
- ARMA model.
- GARCH model.
- Regression analysis.
- Maximum likelihood estimate.

**Tags: **
did you know what accounting USA