AI For Trading:Real World Constraints: Liquidity (71)
Real World Constraints: Liquidity
So far, we've looked at evaluation metrics that assume unlimited liquidity in zero transaction costs.
In other words, we've assumed that we can buy or short the stocks when we want and at the market price without cost or impact to that marketplace.
We also saw that, increasing the number of stocks and the number of trades can improve one evaluation metric, the information ratio. However, when we include real-world constraints and costs, there are reasons why more trades aren't always better.
When trading in the real market, there are constraints due to both liquidity and transaction costs.
Liquidity of a stock refers to how likely one can buy or sell the stock when they want to in the amount that they want to.
A proxy for liquidity is the bid-ask spread. The bid-ask spread is the difference between the price that an investor can buy versus sell immediately in the market.
High liquid stocks that trade in high amounts every day have many shares that are accessible to the public and can be borrowed for short positions. Examples would be large market cap stocks in highly developed stock markets like the US.
The bid-ask spreads in liquid stocks maybe around three to five basis points. For example, if the price at which one can sell a stock is $100 and the price at which one can buy a stock is 100.05,then the bid-ask spread is said to be five basis points.
Low liquidity stocks have fewer accessible shares or are traded on a less active exchange. You may also refer to low liquidity as highly illiquid.
Bid-ask spreads for less liquid stocks maybe 20 to 30 basis points. Remember, a decent Alpha factor might have
an annual return of four percent or 400 basis points. When you think about that, a bid-ask spread of 20 to 30 basis points on a single trade is massive.