If you have an asset that has an illiquidity premium, an optimizer will love it as a choice. An illiquidity premium is a dangerous area for investing. First, are you getting paid enough for illiquid? Second, is there a good way to measure illiquidity? Third, do you really know your liquidity needs?
There are two assets, one liquid and the other illiquid, for consideration. The illiquid asset has a slightly higher yield, a premium for less liquidity. An optimizer will choose the illiquid investment, all else being equal. The illiquid asset will have a smoother volatility. The optimizer will again prefer it versus a liquid alternative. An illiquid investment will not react as quickly to new information and will be less correlated to other assets. The optimizer will again prefer it. An optimizer will love an illiquid asset, and that is a problem.
Now, we have been extreme with our view of a two assets, but the point should be clear. There will be allocation distortions if you do not properly account for illiquidity. Liquidity will never be available when you want it. Illiquidity will never improve from what was sold to you.
So how do you address the illiquidity problem? There are two simple solutions. First, rough-up the data by using a volatility that may be closer to the volatility of the asset class associated with the illiquid asset. This will downgrade the smoothness of the illiquid asset and will also increase the correlation with other assets in the asset class. Second, cut the returns that will represent the risk premia associated with liquidity. An optimizer does not think about liquidity. An investor should account for these factors.