Crystal Ball

Problem

"It is difficult to make predictions, especially about the future."

Policymaking frequently involves predicting the future, but the future is inherently uncertain. Generalist policymakers relying on conventional forecasting methods must aggregate across different sources to make a best guess of the future. It is difficult for them to directly assess evidence and analysis of expert's forecasts as experts are relied upon precisely for their expertise.

Solution

Prediction markets can potentially solve the problem by:

• Giving policymakers a forward-looking view of a future state of the world

• Allow policymakers to pre-emptively stage policy interventions to shift future outcomes

The concept of prediction markets have been endorsed by Nobel-winning economists and works as a mechanism to incorporate different sources of information into a single accurate and canonical view. For more information, read this FAQ.

We have two work tracks.

The first track involves creating a simple demo that allows users to

• Create an account using Singpass and receive 500 units of a virtual currency

• Create Yes/No questions that other users can bet on

• Place bets on questions (via market order or limit order), which would shift the probability via an Automated Market Maker

• Close and resolve the markets, which will give virtual currency to users who bet correctly

• View a leaderboard showing users who have made the most from their bets

Feedback from users:

• “I really like this project! I feel like it’s the most interesting one, it’s such a unique concept!”

• “Do you all gamble a lot?”

• Concept of prediction markets and exact mechanics of how prices and probabilities are generated are confusing

• Concern over whether this would normalise gambling in Singapore

The second work track is to assess the whether policy-relevant questions can be successfully answered using prediction markets by conducting a low-cost proof-of-concept.

We are currently working with agencies to draft a list of such questions that cover a range of different policy areas, different question types, and different levels of uncertainty. When ready, these these questions will be posted on an existing prediction market platform with a thriving user-base and track record of accuracy.

Who We Are

• Zi Xiang Tan

• Gautam Manek

• Stanley Nguyen

• Ryan Tan

• Kenneth Tan