Methodology Weekly activity index

The WAI is a weekly index designed to measure real economic activity in Germany in a timely manner (Eraslan and Götz, 2020). It therefore cannot be equated with quarterly GDP. Similar to an index published by the Federal Reserve Bank of New York (Lewis et al., 2020), it is calculated as a common component from a number of high-frequency indicators that are quickly available and cover various economic sectors. In addition, monthly industrial output and the most recently available quarterly gross domestic product (GDP) feed into the WAI. The WAI and the GDP growth estimate derived from it supply timely information on developments in economic activity in Germany. This is of particular interest at the moment, given the dramatic economic impact of the coronavirus pandemic and the measures taken to contain it.

Construction and interpretation

First, 13-week averages for the high-frequency indicators (available on a weekly basis) are calculated. In a next step, 13-week growth rates are computed from these averages. The WAI is then calculated by means of a principal component analysis using the expectation maximisation algorithm (Dempster et al., 1977; Stock and Watson, 2002) as a weekly factor of the mixed-frequency dataset (weekly indicators, monthly industrial output and quarterly GDP).

The WAI fluctuates around its long-term mean, which, due to the construction of the index (as a result of the standardisation of the data), is zero. As a result, positive values indicate above average growth in real economic activity, while negative values signal a below average increase or a decline in economic output. The WAI provides the trend-adjusted growth rate of economic activity over the past 13-week average (e.g. week 14 to week 26) compared with the average of the preceding 13-week period (e.g. week 1 to week 13). With each WAI update, the periods under analysis shift by one week. Its values may thus be interpreted as rolling 13-week growth rates. As a year consists of 52 weeks (rounded; if it starts or ends on a Thursday, it consists of 53), the values of the WAI at the end of a given quarter can be viewed in approximate terms as the quarter-on-quarter rate of change. A drop (increase) in the WAI indicates a deterioration (improvement) in trend-adjusted activity over the past 13 weeks in comparison with the preceding 13 weeks.


Alongside monthly industrial output and quarterly GDP, the WAI consists of nine high-frequency indicators which are recorded on a weekly basis and quickly available (see table). The daily observations of the respective indicators are aggregated to the weekly frequency by taking their average. One key selection criterion for including an indicator in the index is that it should display sufficient explanatory power in relation to economic activity. Furthermore, as many different economic sectors as possible are to be covered. Thus, the high-frequency indicators “electricity” and “toll” (road charge) capture the production sector and trade, respectively, while the “flights” indicator creates a point of reference for global activity. The “G-unemployment”, “G-short-time work” and “G-state support” variables are derived from Google search queries. The former two relate to the labour market, the latter captures the extent to which state support measures are discussed. The indicators “pedestrian frequency” and “credit card payments” capture parts of consumer behaviour. The “air pollution” variable serves as a metric for the mobility sector. As some indicators display a marked seasonal or calendar pattern, they undergo seasonal and calendar adjustment in advance.




Realised electricity consumption
(excluding industrial grids and producers’ own use, in MWh; source of unadjusted figures: Federal Network Agency (Bundesnetzagentur), 2020).


Daily truck toll mileage index
(Lkw-Maut-Fahrleistungsindex) (source of unadjusted figures: Federal Office for Goods Transport (Bundesamt für Güterverkehr), Federal Statistical Office, 2020).


Worldwide number of flights
(comprises inter alia passenger and cargo flights; source of unadjusted figures: Flightradar24, 2020).


Google search term “unemployment”
(relative search frequency; source of unadjusted figures: Google Trends, 2020).

G-short-time work

Google search term “short-time work”
(relative search frequency; source of unadjusted figures: Google Trends, 2020).

G-state support

Google search term “state support”
(relative search frequency; source of unadjusted figures: Google Trends, 2020).

Pedestrian frequency

Number of passers-by on 21 shopping streets
(Source of unadjusted figures: Hystreet, 2020 obtained from the Federal Statistical Office).

Air pollution

Nitrogen dioxide (NO2) concentration in the air (average of all available measuring stations; source of unadjusted figures: German Environment Agency, 2021).

Credit card payments

Payments based on credit card transaction data in Germany (source of unadjusted figures: Fable Data, 2021)

Further information

The “cash” indicator (withdrawals in euro) that was still part of the earlier version of the model (Deutsche Bundesbank, 2020) has since been replaced by the “pedestrian frequency” indicator. Moreover, two indicators were added during the update on 8 June: Consumer confidence – Index of Current Conditions (ICC) and G-state support. Furthermore, the WAI is since then based on rolling 13-week growth rates instead of 12-week growth rates. Note that these three adaptions lead to a marked downward revision of the index. The source of data for the indicator air pollution was changed from the European Environmental Agency to the German Environment Agency on 11 January 2021. From 8 March 2021, we use 7-day averages (instead of 7-day weighted averages) for the indicator consumer confidence – ICC. Credit card payments were added as a new indicator during the update on 10 May 2021. The indicator consumer confidence – ICC has been removed from the dataset during the update on 1 November 2021. From 16 January 2023, the indicator pedestrian frequency covers shopping streets in 21 instead of 8 cities. It should also be noted that the indicators used may be subject to data revisions which can also result in revisions of past WAI values in the weekly updates.