The Statistischen Bundesamt, also known as Destatis, is the Federal Statistical Office of Germany. It is the institution to contact first for official data on society, economy, environment and the state. Destatis is represented in international statistical bodies of EU, OECD and UN and implements technical assistance projects in official statistics outside the EU. The  Deutsche Bundesbank has a strong interest as a user in federal statistics on the real economy and on prices as part of its monetary policy and financial stability tasks. Destatis has 2404 employees of which 57% are women and 38% of them are managers. We had an interview with Katja Wilken, she is responsable for the Department Population, Finance and Taxes.

The simple answer is: because statistics reflect life. Or to put it more precisely: because good statistics create an accurate picture of social reality in the first place. This can then be used as a basis for fact-based discussion and political decision-making. At the Federal Statistical Office of Germany, we produce some 380 sets of statistics and we are committed to the highest quality standards. Our statistics range from the number of births and the economy to social matters or the equipment of households with smartphones, that is, almost anything that constitutes our social life. We provide facts instead of fake news, we offer real value added for an informed public, for evidence-based policies and, finally, for the development of informed opinions in a democracy. This is what makes working with statistics interesting for me, and it motivates me every day.

Researching new methods of data collection and data processing is one of our focuses. We use machine learning more and more often for automated data plausibility checks and we also use web scraping to get data from the internet. Of course, this is based on complex and "learning" algorithms. Web scraping allows us, for example, to cover dynamic pricing in online trade and to integrate it in calculating the inflation rate. Especially in the lockdown months of March and April, web scraping helped us a lot because it was impossible for some time to collect prices directly in shops.

We also do research on other digital data sources, such as satellite images or mobile phone data. To evaluate such data, we need machine learning and artificial intelligence. By applying these "new" technologies, we wish to process large data quantities in a more timely and more varied manner, so that we can create an even more precise picture of economic and social developments. So, yes: algorithms help us to provide even better data and, consequently, to provide the basis for fact-based political decisions. Of course, we also must make efforts to recruit "digital talents" and to be an attractive employer for IT experts.

Due to the coronavirus pandemic, we have faced particular challenges, which we have mastered well so far: On the one hand, we had to ensure in this unprecedented situation that we can provide our existing range of statistics, while maintaining their usual quality. And we have coped remarkably well. One reason is that the enterprises have continued to provide us with a good data basis despite the crisis. On the other hand, there has never been more demand for highly up-to-date statistics that are also accurate. Just think of political attempts to set up tailored business support programmes or of the public interest in possible excess mortality caused by the coronavirus.

We have implemented nowcasts in addition to existing statistics and have rapidly set up entirely new sets of statistics which are based especially on new digital data sources and are still experimental in some cases. Here is an example: for a number of years already, we have had an indicator showing the truck mileage on motorways, based on toll data. That indicator gives a remarkably good picture of industrial production. To put it simply: when industrial production is running well, there are many trucks on the road, transporting goods. When the industry is doing badly, there are fewer trucks on the road. The advantage this indicator has over others is that it is available very early - we have it after just a few days instead of several weeks after a given point in time. In the crisis, we have changed the base of indicators from monthly to daily, so that now we can show very accurately what happens on the road and also what happens in the economy.

Let me give you another example: by using scanner data, we have been able to show very accurately, and very early, what happens at the supermarket checkout. We have seen, for example, that the demand for disinfectants and toilet tissue shot up when the crisis began and we have seen how it developed during the first weeks of the crisis. One effect of the pandemic is that data produced by well-established methods have been published even more rapidly than would have been the case under normal conditions. The crisis has enormously accelerated these processes. Of course, this is only possible with know how of our colleagues and especially their teamwork and commitment.

If there is anything positive for us about the pandemic, then it is this: we have learned that we will have to react more rapidly and even more flexibly to new data needs in the future. And we have seen what we are able to do. This makes me confident about our future as the leading information provider in Germany.

The gross domestic product is not a comprehensive measure of well-being. The GDP is limited to the economy and it measures a national economy's real economic performance that can be measured in prices - it's as simple as that. But well-being, or the quality of life, comprises much more than that. For example, there is the state of the environment - and this is where the ecological costs of doing business come in.

Other aspects are equal opportunities in education and a fair income distribution. All this means that looking at the GDP is not sufficient if we want to measure well-being. Everyone wants the "one" measure of well-being, but so far no-one has been able to develop it. To get a better picture, we can look at other sets of indicators beyond the GDP. Indicators such as the UN Sustainable Development Goals help us to widen our view to include the social and ecological dimensions of well-being. For a number of years, the Federal Statistical Office has monitored the SDGs in Germany and has published the data - and this is really a good complement to the purely economic measurement of well-being. In addition, we cooperate with statisticians worldwide to develop a measure beyond the current GDP that includes unpaid housework, the distribution of income and consumption, saving and assets, and health and social matters. But I have to admit that this is a thing of the future.

To remain relevant, our statistics will have to be of great benefit also in the 21st century. This means that they have to be up to date, they have to represent reality very accurately and, above all, it must be possible for users to evaluate them comprehensively. We are therefore making great efforts to develop not only new digital data sources but also new data hubs and dashboards. A general aspect in all this is the open data idea. All results should be available in open formats, and machine evaluation should be possible. Obviously, we also have to close data gaps, become more flexible and reduce the burden on respondents. 

No one wants to complete long questionnaires, and no one wants bureaucracy, but everyone wants good statistics. For us this means that we have to use more often existing data from administrative registers. And this means that registers that are decentralised, because of the federal structure of Germany, have to be modernised and linked up. This requires legal amendments, among other things. So you see - we are working in many complex areas of action. But I think we are well prepared for the future. What is exciting in the digital era is the permanent disruption and the enormous technological opportunities. This is what fascinates me as a statistician. At Destatis, we want to use that potential to enhance the benefit of our statistics.

destatis.de

Afbeelding gebouw en Katja Wilken © Destatis