I don't think this conversation can really proceed in a sensible manner without the introduction of some numbers.

Retail brokers typically charge a commission per trade. If you want to buy many different securities ('cuz diversification), and you don't have a whole lot of money to invest in the first place, you're going to end up paying a large percentage of your initial investment in commission. The exact amount will depend on your broker and how much money you had to begin with, and how many stocks you're diversifying among. That means you might be starting from a fairly deep hole to have to dig yourself out of before you're truly earning a positive return.

By contrast, with ETFs you only have to pay commission on no more than a handful of trades to get a well-diversified initial investment laid out. Your expected earnings rate might be fractionally lower, but since you're starting from a much shallower hole, you might have a decent head start compared to buying a well-diversified portfolio individual stocks.

So then you've got two theoretical curves describing how your wealth might grow, and whether one is more favorable than the other depends on whether those two lines are likely to intersect at a point that comes before your investment horizon.

Then you can throw in still more complications that, IMO, can make ETFs still look quite a bit more favorable than stocks for most investors. One is that many retail brokerages (Vanguard, for example) will let you buy a selection of ETFs for zero commission. That's a gift that keeps on giving, since it dramatically reduces the cost of making smaller investments more frequently.

Another is that picking stocks is a time-consuming process - you have to spend time learning how to do it, and then you have to spend time researching stocks. If you agree that time is money, then you should probably be including some estimate of the value of your time into the formula. You want your expected returns from manually selecting a portfolio to be great enough to justify your time. Which is yet another calculation that is going to be heavily influenced by individual factors, particularly how much money you have to play with, what your current earnings are, and most importantly, whether or not you think it's fun to pick stocks.

> Another is that picking stocks is a time-consuming process

Just buy them all. Or pick them randomly. Let me remind you that this thread started with an article about the guy who wrote about the blindfolded monkeys.

It would cost $2500 in fees for me to buy a single share of every stock in the S&P500. Compared with $0 + some negligible MER to buy an ETF with the same money.

If you want to buy every stock in the S&P500, any reason not to use Robin Hood with their $0 transaction fee?

It's not available in Canada.

My guess is that he was being hyperbolic. But even if he weren't, you can still get something that has the same expected return over time as picking individual stocks at random for a much lower up-front price by choosing a total stock market ETF. Remember that he was writing that comment at a time when the latter option didn't exist.

So your strategy isn't, "buy the whole market", it's "buy the market as it existed in $current-year", which misses all the new stocks that come out. (Also it assumes you have a fixed amount to invest and never want to invest more)

I don't get what you mean. You can buy and hold, and keep buying.

But then it's not a one time-fee. More generally, your strategy only works if you're investing a large amount every time, it doesn't scale down. You can't put 10k into 5000 stocks, and even 500 stocks would be difficult.

> But then it's not a one time-fee.

I meant one time per line, not per portfolio.

> it doesn't scale down.

It scales down just fine if you're willing to accept that you're going to pay a lot in broker fees, with the only consolation that you'll pay them only once per line in your portfolio.

If you have say 500 stocks and pay like $10 in broker fees per stock, in the end you'll pay $5000. That's not the end of the world for a long term investment.

At a 0.05% expense ratio, you need to hold $250k for 40 years to pay $5000, not to mention the extra fees if you want to rebalance as the market changes. It doesn't seem to make sense for lower-end investors, which are many of us.

> If you have say 500 stocks and pay like $10 in broker fees per stock, in the end you'll pay $5000. That's not the end of the world for a long term investment.

If you only buy each stock one time. Which is absurd if you want a balanced portfolio you would need to buy a little bit of all 500 stocks a lot of times. A ETF lets you do this cheaply.