The new experimental early indicator of short-term development in the non-financial business economy, which is based on advance turnover tax returns, permits early information on turnover development in the non-financial business economy to be provided before the official survey results broken down by economic sector will be available. The data are based on the advance turnover tax returns which are transmitted to the statistical offices in accordance with the Act on the Use of Administrative Data (German: Verwaltungsdatenverwendungsgesetz). Evaluating the data in this form is something new and therefore of an experimental nature.
The nominal, calendar and seasonally adjusted turnover index for the "Total non-financial business economy" aggregate (Sections B to N without L and K according to the German Classification of Economic Activities, 2008 edition - WZ 2008) will from now on be updated monthly and published here on the EXSTAT - Experimental statistics page of the Federal Statistical Office's website. The results on development of turnover are published in our press releases.
According to the statutory deadlines, advance turnover tax returns must be submitted to the fiscal authorities not later than ten days after the end of the reference period (months or quarter). A permanent extension of the deadline - this is requested by approx. 70% of those obliged to submit advance returns (with a turnover share of more than 80%) - extends the submission deadline by one month. For evaluations of the current reference month with a timeliness of less than 30 days after the end of the month, the data are therefore still quite incomplete. Generally, data are available for about 20-25% of turnover in the current month.
Only when most of the advance returns from enterprises with a permanent extension are available in the following month can the data be evaluated with an almost complete data basis and a timeliness of 60 days. Within the framework of mixed models, these data are regularly used to supplement short-term statistics on trade, construction and services. Using mixed models means that large units are surveyed in primary surveys, while information on small and medium-sized enterprises is obtained from administrative data, like those of the advance turnover tax returns.
To compile the early indicator, the advance turnover tax returns are for the first time evaluated across the entire non-financial business economy. This is an evaluation of purely administrative data, to which no information from primary surveys is added, not even for previous months.
The turnover index is calculated from the individual month-on-month rates of change, which are determined according to what is called the pairing concept: only those enterprises are included in a rate of change with their data for which values are available for both the reference and the previous month. For the preceding months, the rates of change calculated with complete data are used. As the data basis is incomplete for determining the current month-on-month rate of change, the available turnover data are extrapolated in a breakdown by economic sector. This is an extrapolation with auxiliary information, in which the turnover values available for an economic sector are adjusted to the turnover level of the preceding month with complete data, using an extrapolation factor. If, for example, the enterprises for which current data are available accounted for 20% of the total turnover of a given economic sector in the previous month, the grossing factor of this sector will be 5 (inverse of 20%). This approach ensures that the weight of each economic sector represented in the total result is in line with the situation in the previous month, irrespective of the proportion of current turnover reported by early respondents in the economic sector.
In addition, calendar and seasonal adjustment is required to enable interpretation of the data for purposes of short-term economic statistics and ensure comparability with other well-established short-term indicators of the Federal Statistical Office. The seasonal adjustment method used for this purpose is X13 in JDemetra+.
The method was originally intended to also correct systematic biases between the incomplete current and the complete data using calendar, seasonal and trend factors. In terms of the need for revision, this showed good results for the reference months before the outbreak of the coronavirus pandemic. Due to the continuing special circumstances caused by the coronavirus pandemic, it has however turned out to be more appropriate to do without this estimation based on past values at the current end. As of reference month October 2020, the procedure has therefore been modified; the current month-on-month rate of change of the early indicator is now produced based on the incomplete data without using a correcting estimation.