top of page

Portfolio Market Making Strategy

Writer's picture: helen ristovhelen ristov

The trading execution is based on portfolio risk minimization using principal components analysis on the existing asset classes in the portfolio.

Given a current portfolio comprised of n asset classes X1…. Xn we can compute the volatility of the portfolio, given the corresponding net positions p1 …. pn. The underlying risk of the portfolio on those positions is given below:


σP= √(P*Σ*PT )

where P is the vector of positions and ∑ is the covariance ma

trix among the asset classes.

Given a current portfolio and market state, the portfolio market making strategy determines the quantities, or equivalent positions (EP), of each asset class that would minimize risk:


=

 EPi= Σi /(σi2) * P

From these positions, we create another risk metric, namely equivalent levels (EL) that identifies how many block orders of the same shown quantity (SQ) must be executed to minimize risk using only that contract.


 ELi = EPi/SQ

Henceforth, the strategy is fully automated and calculates the prices for which new orders should be put on and existing orders should be cancelled. These prices are evaluated and updated by creating premium requirements around what we define to be the passive executable fair market value or FMV. A custom algorithm is used to determine the FMV.

Total Price Premium for bid and ask prices are determined by fixed and variable components. The fixed quantity establishes how wide we want to be from FMV and the variable component adjusts with position (risk).

TP = FC + VC where

FC = FP*(Vol)

There are two types of Variable Component equations that can be used:

Fixed Cheat-In/Back-Off : VC(1) = VP*(Vol + EL*CI) -or- VP*(Vol + EL*BO)

Vol-Adjusted Cheat-In/Back-Off: VC(2) = VP*Vol*( ­1 + EL*CI) -or- VP*Vol*(1 + EL*BO)

The strategy is parameterized and developed on a platform that allows the user to adjust certain parameters and make decisions based on the state of the market and portfolio risk. The orders are fulfilled using a custom GUI and the TT platform/infrastructure.

Two strategy execution modes are available depending on whether or not you want to trade outrights. The positions that minimize risk are solved using constrained optimization.

Constrained — The strategy calculates the equivalent positions that minimize risk not allowing outright execution.

Unconstrained — The strategy calculates the equivalent positions that minimize risk allowing outright execution.

Strategy Parameters:

Reference Quantity: An amount that adjusts the shown quantity of the contracts that the trader is willing to put on at any given time.

Fixed Premium (FP): A fixed value that offsets the strategy’s bid-ask prices from the fair market value. This determines how wide we want to be from FMV.

Variable Premium (VP): A value that adjusts the bid-ask prices based on relative position/risk in the portfolio. This value will change with position and should be considered as sensitivity to position.

Cheat-in (CI): Multipliers to variable premium determining the speed at which the strategy will try to remove risk. For example, if we have a long portfolio, we would cheat-in the ask price to get closer to fair market value and sell a trade.

Back-Off (BO): Multipliers to variable premium determining the speed at which the strategy will move away from fair value as it builds an exposure.

Create Premium (CP): Additional Premium required to initiate a new order on a tighter market price in order to avoid flash orders. This should be considered as an indifference zone around new order and cancel order prices.

Queue Holding Offset: A distance from fair value that bid prices and ask prices are set to when we wish to hold queue in the book and preserve place in the market, but we do not wish to get trades.

Terms

Reference Volatility (Vol): Volatility of the contracts in relationship to an anchor contract.

Shown Quantity: The risk adjusted amounts that you are willing put on for each contract

Equivalent Positions: The number of positions that you would add/subtract from a contract in the portfolio to minimize the overall portfolio risk using only that contract.

Equivalent Levels: How many times you would have to execute the shown quantity for a given contract to minimize the overall portfolio risk given only that contract.

Passive Executable Ask/Bid : The fair market price at which a buyer or seller would execute orders immediately to get a reasonable price and not sweep the stack. This is the baseline off of which bid and ask premiums are added.

Create Bid Adjustment (CBA): The adjustment to the market’s passive executable ask fair market value that will determine the price point at which we will create a new bid order. It is calculated using the total premium equations.

Create Ask Adjustment (CAA): The adjustment to the market’s passive executable bid fair market value that will determine the price point at which we will create a new ask order. It is calculated using the total premium equations.

Cancel Ask Adjustment (XAA) : The adjustment to the market’s passive executable bid fair market value that will determine the price point at which we will cancel an existing ask order. This is the outdated ask premium amount.

Cancel Bid Adjustment (XBA): The adjustment to the market’s passive executable ask fair market value that will determine the price point at which we will cancel an existing bid order. This is the outdated bid premium amount.

5 views0 comments

Recent Posts

See All

Comments


bottom of page