Theme 7: Economic Surprises #
Quantamental data surprises are deviations of point-in-time quantamental indicators from predicted values. There are two information events with respect to these surprises: first-print events and pure revision events.
• A first print event is the difference between a quantamental indicator value for a new observation period and its predicted value on a release date. Here, a release date is the day on which the observation period of a quantamental indicator changes. Predictions in a quantamental system are generally based on econometric models and information prior to the release date. Since one can use different models, there can be different first-print surprises. A quantamental data surprise is always defined conditional on a specific model.
• A pure revision event is the change in a quantamental indicator on a non-release date. It arises from revisions of data for observation periods that have already been released. Per default, we assume that all revisions are surprises, i.e., the quantamental system makes no attempt to predict them. Any revisions published on a release date become part of the first print event. A quantamental indicator of surprises presents the values of these two events on the dates they become known. It contains zero values for all dates when neither releases nor revisions occur.
It is essential to differentiate between the quantamental indicator and its underlying data series. A quantamental indicator is typically a transformation of one or more economic time series (and their vintages). Prediction models are always applied to the latest vintage of the underlying data series and produce a future predicted vintage of that series. More specifically, prediction models use symmetric changes, i.e., differences or log differences, to measure activity or stock. For example, suppose annual industrial production growth rate predictions are based on an ARMA(1, 1) model. In that case, the model is estimated and used for the monthly percent changes of a seasonally-adjusted index. Then, the next month’s prediction is added, and the new annual growth rate is calculated on this updated vintage. Thereby, forecasting automatically accounts for base effects.
Quantamental indicators can be based on more than one underlying data series. In this case, the release data of the quantamental indicator for a new observation period is the release date of its main underlying series. For example, for a trade balance-to-GDP ratio, the release of the trade balance determines when the quantamental ratio shifts to a new observation period. Any subsequent release of a GDP for the period will be treated as a revision event. There will only ever be one first-print event per observation period and per quantamental indicator.
