Problem Definition

A simple question to ask is why raise fees in higher volatility?

A few concepts behind this are adverse selection, risk management and surge pricing.

A common example improperly cited is what market makers do in times of volatility.

https://s3-us-west-2.amazonaws.com/secure.notion-static.com/a12e1fd7-5714-45cf-b17b-e15704586bd0/Untitled.png

Variance in traditional financial markets can be scheduled (Non-farm payrolls, Federal Open Market Committee meeting - known unknowns) or unscheduled (Unexpected tariff announcement, Suez canal blocked - unknown unknowns).

Adverse Selection

The clearest case of market makers widening out is just ahead of scheduled variance events since there are often lots of NLP based market takers looking to trade on informational advantage from parsing the relevant headlines or documents just as they are publicly released. This widening period is often very brief and only for a few seconds. In this instance the fee implied by the spread goes up to protect the liquidity provider against highly informed order flow in a burst of volatility.

Risk Management

In unscheduled variance the effect on price is either directional or noisy. In directional movements market makers could widen just one side of the market or not widen at all and simply shift their market about their conception of fair price based on the market and their own positions. In noisy movements market makers also need not widen depending on their competition. No clear direction but lots of opinions tends to lead to a flurry of trading so the volatility brings volume and higher profits. In either case the consideration shifts more towards risk management and how close to market neutral a liquidity provider's portfolio is (assuming they have no opinion on direction). In this instance the spread does not necessarily widen much with volatility. Historically market maker profits largely come in volatile markets but these are usually due to trading opportunities from correcting inefficiencies rather than spread widening.

Surge Pricing

More recently popularized by ride-sharers like Uber is surge pricing where the price is determined algorithmically based on supply and demand. If pool utilization is exceptionally high this indicates low supply or high demand or both. Such cases could warrant a fee increase depending on the competitive market with activity spikes often being short-lived (rain for ride-shares, new yield farm for pools etc). Here the fee increase based on volatility is based on volatility bringing more demand, shrinking supply or both which would need to be tested to justify this.

Deciding which subset of these are actually being done is an important prerequisite for action.

Model Scoping

Model Definition

Heuristic based approach where if a couple risk factors exceed a high threshold raise the fee and then decay it back to a baseline. Highest suggested frequency for an update would be hourly.

Model Factors