Financial risk management in a decentralised setting

🧠 Ideas for the Protocol, Research, and Markets
1

A discussion thread for our research problem centred around decentralised risk management:

When dealing with financial products it is import to accurately estimate risks involved. This requirement is even more pressing in the case of derivatives. While there are many potential sources of risk, some of which are quite esoteric and difficult to quantify, market risk is often the one that is considered the most.

For the purposes of this problem we can define market risk as distribution of prices for a given market at a fixed point in the future. Any form of risk management requires taking a view on what that future price distribution may be.

In a centralised setting it would be chosen by single entity according to well defined set of rules and regulations. In a decentralised setting anyone should be able to put forward their view on it.

As this process is non-trivial and has high potential consequences for the market an adequate incentive scheme must be put in place to encourage views that are most likely to be accurate (and are not deliberately manipulated to extract gains from the market by acting as a market-participant). Downplaying potential downside and exaggerating upside of future market moves might encourage excessive risk-taking and in turn a series of defaults possibly followed by a market crash. An overly conservative prognosis might reduce market activity which is an undesirable outcome from the point of view of the exchange.

1 Like
2

Link to @david’s paper formalising the problem and providing empirical results for a risk model linear in the calibration parameters. Abstract below:

We consider the problem of risk model calibration that is faced by all decentralized derivative exchanges. Financial model calibration is hard for two reasons: firstly it relies on data inputs that can be unreliable, incorrect and in general needing manual cleaning. Secondly, even if perfectly correct data is available the problem typically involves non-convex minimization resulting in local minima that are moreover highly dependent on small perturbations to input. Thus even honest parties won’t necessarily produce the same parameters. On a decentralized exchange multiple parties need to agree on the correct calibration. Moreover, malicious actors may benefit in providing calibration parameters that benefit their trading, in case they can convince others that their calibration is the right one. Effectively we have a problem of trying to achieve consensus in continuum.

We propose a phenomenological model for the problem. We analyse this in the framework of stochastic differential games and we show that a Nash equilibrium exists. We present empirical results for simple situations that arise when the risk model is assumed to be a linear function of calibration parameters.