Measuring Output Gap Uncertainty – Applied Forecasting

This paper assesses the utility of qualitative expectational survey data at the firm-level in terms of both their ability to anticipate firmsâ subsequent retrospective, but qualitative, reports of their performance but also these same firmsâ quantitative answers. The assessment requires access to a unique panel dataset which matches firmsâ responses to a leading qualitative tendency survey conducted by the Confederation of British Industry with these same firmsâ quantitative replies to a different survey carried out by the Office for National Statistics. We employ nonparametric tests of the so-called âbest-case scenarioâ and introduce a weaker test for the coherence between these two surveys and test whether the qualitative data contain a(ny) signal about the quantitative data. We find that while firmsâ qualitative expectations are âbest-caseâ predictions of their qualitative assessment of their output growth they do not contain a signal about the quantitative data. But we can reject the null hypothesis of noise for the retrospective qualitative data. We discuss this apparent paradox and suggest that qualitative business survey data are more useful for nowcasting than forecasting.