"If there is a 25% immediate cost of selling and estimated 5yr...

"If there is a 25% immediate cost of selling and estimated 5yr return of 10%pa you can backsolve the required return for a high conviction investment (eg 80% probability of success) and make calculated decisions regarding optimal capital allocation. Considering @madamswer that you emphasize capital allocation discipline in your businesses I am surprised you don’t have the same requirments for your own investment process. I contend that if you are not systematically doing some variation of the above you will eventually drift downwards in terms of overall portfolio return."

@mal85 ,

From a pure arithmetic standpoint, you are right, but there are certain practical limitations to following such a mechanistic approach.

For starters, it assumes a marketplace of sufficient liquidity and an adequate number of viable investment alternatives, neither which is the case.

Also, it assumes that 25% is the cost of selling. When markets are as fully-valued as they are today, the capital gains liability drag could easily be a great deal more than that.

And, ironically, the more stretched the valuation - which is when one would be most inclined to sell - the greater the tax drag and therefore the greater the return required from whatever new stocks the liberated capital is invested.

Meaning that - when markets are expensive, like they are now - by selling my expensive stocks, I'm incurring a great tax impost and, as a result, I would need to generate a higher return from whatever new stock(s) I buy... at a time when most other stocks are expensive to begin with.

It's a veritable Gordian knot, and an example of why this investing lark is not an exact science.