Bundesbank: Unsupervised Risk Monitoring

Guest Blog Deutsche Bundesbank — The Deutsche Bundesbank is the independent central bank of the Federal Republic of Germany. Together with the other national central banks in the euro zone and the European Central Bank, it forms the Eurosystem.

3 min readNov 12, 2020
Copyright: Deutsche Bundesbank

Unsupervised Risk Monitoring — a macroeconomic perspective

In August we launched the Bundesbank Innovation Challenge — looking for innovative risk monitoring solutions focusing on AI and Big Data. We now want to share some further insights about our interests and potential solutions we are interested in.

The Bundesbank

At the Bundesbank we combine many tasks: We supervise banks, we ensure smooth payment transactions, we monitor risks to financial stability and much more — but monetary policy remains our primary mandate. Together with our partners in the Eurosystem, our primary objective is to maintain price stability in the eurozone. However, price developments depend on many factors. In our Economics Department, we analyse many different macroeconomic variables. Through in-depth analyses of the German economy, we monitor the current economic situation and produce forecasts about future developments. These analyses support our monetary policy tasks in the Eurosystem.

Risk Monitoring Solutions

As one can imagine, the development of the economy depends on many things. Different events can trigger changes or turning points. However, it does not take a pandemic such as Covid-19 to have a significant impact on macro-economic forecasts or observable trends. Legislative changes, political decisions, elections, trade tensions or consumer sentiment are just a few examples that can play a role. Forecasts of macro-economic factors are therefore complex and based on a wide variety of data sources. This is where the Innovation Challenge comes into play: technologies can support and improve such forecasts. Here we hope for creative and exciting input from the startup side.

Considering the above, we are hoping the Innovation Challenge could contribute to two especially challenging issues.

First Challenge

The first challenge is the development of productivity. Both total factor productivity, i.e. that part of the output that is not directly attributable to the quantitative use of production factors, and the productivity of the individual input factors, i.e. labour productivity or capital productivity, play a central role for longer-term economic growth.

A large number of factors influence the development of productivity. New developments or certain events can have a significant impact on the productivity path. Examples are the outsourcing of labor-intensive processes in the industry to Central and Eastern European countries in the 1990s and the labor market reforms in the first decade of this century. Currently, digitization is seen as a possible driver of future productivity enhancements.

We are looking for possibilities to detect structural changes in productivity and to predict trends in long-term productivity paths. The input data is up to your imagination — everything that you think could predict such changes. Ideally, the output is able to identify drivers behind the predicted changes.

Second Challenge

The second challenge is about early detection of risks for the economic cycle. Macroeconomic developments do not usually occur evenly over time. Instead, there are different phases. In some phases, the economy grows at an above-average rate, in other phases at a slower rate, and there are phases in which economic output declines.

For the analysis of the economic cycle and monetary policy, it is of particular interest to anticipate and predict strong changes in the pace of growth as early as possible, especially cyclical turning points or recessions. We already use different models for this purpose (e.g. Eraslan/Nöller 2020, Eraslan/Götz 2020). However, it would be insightful for us to get to know more recent, also innovative approaches, both in terms of methods and in terms of the data used.Keep in mind:

What we as the Bundesbank want to know is, what is the probability for certain events (such as a recession or an economic downturn) when looking at GDP or industry production?

Additionally, which data could be helpful to identify risks to the economic cycle?

Let’s get to it!

We are curious about the ideas of the participating startups, their approaches and experiences with Predictive Analytics, Anomaly Detection and Big Data! We are excited to get to know them and to hear about their solutions!

For interested parties who want to get to know the startups and their solutions to the challenges, the DemoDay for the Bundesbank Innovation Challenge will be live-streamed. Get your free ticket now!




We build startup ecosystems. As a digital innovation hub in Frankfurt (Germany), we accelerate entrepreneurship.