backtesting vs Liquidity

[Rufolo]

Good morning
i’ve recently made a backtesting with our universe of 400 latam stocks, using PBV and ROE 12 month forward, from 2002 to 2013. I’ve notice that the portfolio simulation has constructed, first years, many portfolios with lots of stocks so iliquid. I’d like to know if you think there is a constraint or a way to avoid these kind of companies.

Thanks a lot.

Kind regards,

 

[ShaktiRathore]

Hi
Liquidity is preferable while doing back-testing. use the formula for Liquidity Adjusted Var while doing back-testing the Var will always increase due to liquidity as we are unsure of the future risk if liquidating higher the liquidity better chance to liquidate positions but illiquidity creates illiquid positions so its difficult and costly to liquidate during heavy loss or downtime but in normal var calculation we always assume no problem in liquidating loss positions which is not always the case especially with the illiquid stocks. You can always include the liquidity premium by adding liquidity component to var,LVaR = -mean + volatility*sigma + LC; i.e., +LC increases a positive VaR. If you are including this illiquid stocks with liquid ones than try to make the adjustment otherwise its preferably to leave aside illiquid stocks. you need to have data for spreads and other proxies for illiquidity to make the adjustments.
thanks