The problem with the above, though, is that you’re only looking at half the ledger. You’re running the same risk - you’re just taking the other side of it, and outside your investment account statements.
Remember, when you rent instead of buy, you don’t just get the asset sitting in your brokerage account for 40 years. You also then have to pay rent for those 40 years. This is what ML was pointing out on the buy side - you obviously aren’t making a net 10% return on your house because you have to pay your mortgage. You’re not making 10% return in the rent scenario, either, because you have to pay your rent if you don’t buy the house.
So it’s certainly true that if you buy instead of rent, you run a risk that the cost of housing will appreciate by less than that 4% - and then you’d have been better off renting. But if you rent instead of buy, you run the risk that the cost of housing will appreciate by more than 4%. Your rent goes up faster in that scenario. The assumption that your rental payments are equal to (or less than) the cost of servicing the mortgage breaks down. So by the time you get to the end of the 40 years, the returns in your brokerage account get eaten up by the increase in rent.
To use your Houston example, rents only went up by about 1.7% per year - the same as the value of the house. But in a market where residential real estate is increasing at 4% per year, the rents will roughly track it. So in your life history, rents only went from $400 to $800 per month - so the renter would have only paid about $300K over that 40 years compared to the $690K asset. But in a market where housing costs increase 4% per year (the national average), rents would have increased from $400 to about $2000 - and you would have paid closer to $600K in rent, all-but-offsetting the value of your investment.