EN
The main aim of the paper is to research usefulness of business survey results of IRG SGH for forecasting of manufacturing production yearly index (YoY). For this purposes there were used single equation regressions. The regressions use as a explanatory variables business survey indexes with different possible leads. Paper uses 8 questions from the IRG SGH industrial survey in the perception and expectation form and also general business indicator. All model are investigated for raw, seasonally adjusted and smoothed time series. In addition qualitative models are compared with autoregression model. The general conclusion of the research are: models for seasonally adjusted and smoothed series have better forecasting properties; qualitative models have better forecasting properties than autoregression model; qualitative models allow for two-months ahead forecasts; the models for balances have comparable forecasting properties or even better than models based on fractions from particular questions