Price prediction is extremely difficult because, price fluctuations are small and are secondary to liquidity fluctuation. A question arises whether liquidity deficit can be traded directly. If we accept that liquidity deficit is an entity of the same nature as volatility then the answer is yes, and liquidity deficit can be traded through some kind of derivative instruments. Let us illustrate the approach on a simple case – options trading. Whatever option model is used, the key element of it is implied volatility. Implied volatility trading strategy can be implemented through trading some delta–neutral “synthetic asset”, built e.g. as long–short pairs of a call on an asset and an asset itself, call–put pairs or similar “delta–neutral vehicles”. Delta neutral is a portfolio strategy consisting of multiple positions with offsetting positive and negative deltas so that the overall delta of the assets in questions totals zero. A delta-neutral portfolio balances the response to market movements for a certain range to bring the net change of the position to zero. Delta measures how much an option’s price changes when the underlying security’s price changes. Optimal implementation of such “synthetic asset” depends on commissions, liquidity available, exchange access, etc. and varies from fund to fund. Assume we have built such delta–neutral instrument, the price of which depend on volatility only. How to trade it? We have the same two requirements: 1) Avoid catastrophic P&L drain and 2) Predict future value of volatility (forward volatility). Now, when trading delta–neutral strategy, this matches exactly our theory and trading algorithm becomes this.
- If for underlying asset we have (execution flow at time t = 0) I0 < IIL (liquidity deficit) then enter “long volatility” position for “delta–neutral” synthetic asset. This enter condition means that if current execution flow is low – future value of it will be high. If at current price, the value of I0 is low – the price would change to increase future I.
- If for underlying asset we have (execution flow at time t = 0) I0 > IIH then close existing “long volatility” position for “delta–neutral” synthetic asset. At high I0 future value of I cannot be determined, it can either go down (typically) or increase even more (much more seldom, but just few such events sufficient to incur catastrophic P&L drain). According to main concept of P&L trading strategy, one should have zero position during market uncertainty.
The reason why this strategy is expected to be profitable is that experiments show that implied volatility is very much price fluctuation–dependent, and execution flow spikes I0 > IIH in underlying asset typically lead to substantial price move of it and then implied volatility increase for “synthetic asset”. This strategy is a typical “buy low volatility”, then “sell high volatility”. The key difference from regular case is that, instead of price volatility, liquidity deficit is used as a proxy to forward volatility. The described strategy never goes short volatility, so catastrophic P&L drain is unlikely. In addition to that, actual trading implementation requires the use of “delta–neutral” synthetic asset, what incurs substantial costs on commissions and execution, and thus actual P&L is difficult to estimate without existing setup for high–frequency option trading.