Most tokenomics policies today (e.g. inflation rate, burning rate) are either coded statically or rely on DAO governance for adjustments. This approach ensures decentralization, but the lack of real-time response mechanisms also make the tokenomics vulnerable to the impact of external shocks, such as a large price movements and hacks.

How do we design tokenomics that is robust against these shocks, while preserving decentralization? The bold yet rational answer is: let a bot run the economy.

In this research project, my aim is to build an autonomous algorithm that regulates a blockchain economy with two key parameters: staking rate and burning rate. In particular, I apply a modeling tool from modern macroeconomic theory called DSGE, which is the gold standard used by central banks worldwide to determine key economic parameters like interest rates. Cutting-edge data science tools such as machine learning and causal inference models will be applied in the calibration process. This research will provide guidance for the new generation of protocol tokenomics.

About the researcher

My name is Ming and I am a graduate student and researcher at UChicago studying computational economics. I've worked as a research fellow at Nethermind and DeFi research analyst at Polygon. You can learn more about me and my previous work at https://sites.google.com/view/mingxuanhe

Improvements on existing approaches

The idea of optimizing burning and staking using dynamic modeling tools has seen growing exploration recently (e.g. @urosnoetic's post about BME in the previous round of TE grants, and multiple papers from the economics/finance literature in 2022/2023). My novelty is to develop a fully testable, calibratable model that also addresses uncertainty and lack of information. I also address key concerns such as algorithmic bias and systemic economic inequalities.

Current state

I have formalize the baseline model (Biais et al, 2020) in the context of tokenomics, and I'm currently in the phase of building the full model. Presentation slides can be found under the project github repo. The slides summarizes the current state of the research project, include a detailed introduction for the project, literature review, some technical components, and next steps.

Expected outcome

The expected result is a complete MA thesis paper (15-30 pages) which will go under the review of a graduate research committee at University of Chicago Kenneth C. Griffin Department of Economics. The end goal is to publish the paper in a peer-reviewed journal or conference. Before publication, I will open source the working paper and any relevant code on Github. The project is expected to take 4-5 months, with an upper limit of 8 months.

Funding Usage

As an international student in the US, it is extremely difficult to receive academic scholarships and research funding because of my foreign status. That's why I turned to this community of fellow builders for help. Funding from this grant would help me cover part of my advanced studies in computational economics as well as some paid research tools. Here's a plan on how the funding will be used on this project:

• Data: will try to use public data for initial stages, but later stages likely requires subscription-based APIs or databases (~$50/month * 5 months)
• Computational resources: I will need to use credit-based computing services such as AWS or the Midway Cluster for running large-scale computation and machine learning models. An example can be found at https://github.com/mingxuan-he/eth-fee-time-series (~$80/month * 5 months)
• Journal/conference submission & traveling expenses (~$1000)
• Total: ~$1650 Any additional funding will be directed towards creating more accessible community content such as explainers, blog posts, and open-source widgets.