Forecasting developments in digital policy

As in every other area, forecasts help us imagine what the future will look like, and what to expect. Over the years, we have been providing annual predictions in digital policy – see our predictions for 2015, 2014, 2013, 2012, and 2011.

Our foresight methodology has developed gradually over the past years. The current version of Diplo’s forecasting model, which we have used for this year’s predictions, includes the following inputs:

1. Expert knowledge in digital policy. Since the digital policy field is quite predictable, it is very often sufficient to have expert analysis for a reliable prediction.
2. The knowledge of many, including Diplo’s lecturers, students and Alumni. The most valuable insights come from those who are close enough to have a basic understanding of digital policy but distant enough to avoid tunnel vision.
3. Data-mining, which provides inputs on emerging patterns which are difficult to detect in day-to-day dealings with digital policy issues. For example, ‘Internet governance’ was used as an umbrella term for all aspects of digital policy for many years since WSIS in 2005. However, in 2013, the first signs of use of the term ‘cybersecurity’ started appearing in parallel to the use of ‘Internet governance’. This trend was identified in academic articles (mainly from authors dealing with cybersecurity), and later on shifting to semi-policy processes. The WSIS+10 resolution deals with Internet governance and cybersecurity separately.

With more data on digital policy available online, data-mining will be more important and useful in identifying emerging policy patterns.

Any forecasting approach must have strong in-built feedback loops. For example, we revisit our predictions at the end of each year and see how accurate our forecast was (as we did in our 2016 predictions). Sometimes, this exercise can identify biases which we were not aware of. Another important aspect is that forecasters need to revisit their predictions in view of new information gathered, such as through data-mining. This leads us to the challenging exercise of putting all the inputs together.

In 2016, we plan to introduce further innovation to our prediction methodology, including the use of surveys to will gather input from various communities, with the aim of combining statistics and psychology.