For most investors, ETFs are the better default. A single broad-market ETF such as VTI holds 3,639 companies and charges 0.03% — about $3 a year on a $10,000 position — and it delivers the market’s return without requiring you to be right about any one company. Individual stocks offer unlimited upside and no expense ratio, but the historical distribution is punishing: research by Hendrik Bessembinder found that four of every seven US common stocks since 1926 produced a lifetime return below one-month Treasury bills, and just 4% of listed companies account for the entire net wealth creation of the US stock market. The sensible structure for most people is core-and-satellite: put the bulk of the money in one or two broad ETFs, and cap individual stock picks at a small, deliberate slice.
This is the comparison that sits underneath almost every beginner portfolio question, and it is usually argued with slogans rather than numbers. So here are the numbers — on cost, on diversification, on what individual stocks have actually returned, and on the handful of things a stock genuinely does better than a fund.
The Core Difference
An ETF is a single security that holds a basket of other securities. When you buy one share of VTI, you own a sliver of every investable public company in the United States. A stock is a claim on one business. That is the entire structural distinction, and everything else — the risk profile, the fee, the amount of work involved — follows from it.
The consequence people underrate is what happens when something goes wrong. If a company inside VTI commits fraud and goes to zero, your loss is a rounding error, because the index simply drops it and moves on. If that company was the one stock you owned, your position is gone. Diversification does not increase your expected return; it removes the outcomes where you get wiped out by a single event you could not have predicted. (If you are still sorting out what an ETF is relative to its neighbours, start with our explainer on ETF vs index fund — they are not opposites.)
Head to Head
| Factor | Broad-market ETF (e.g. VTI) | Individual stock |
|---|---|---|
| Companies owned | 3,639 in one purchase | 1 |
| Annual fee | 0.03% ($3 per $10,000) | None |
| Trading commission | $0 at major US brokers | $0 at major US brokers |
| Company-specific risk | Effectively eliminated | Fully concentrated |
| Best realistic outcome | The market's return | Many multiples of the market |
| Worst realistic outcome | The market's drawdown | Total loss (−100%) |
| Research required | Almost none | Ongoing, per company |
| Tax-loss harvesting | Only if the whole fund is down | Lot-level, per position |
| Dividend control | Whatever the index pays | You choose the payers |
| Historical base rate | Market return, reliably | 57% trailed T-bills (1926–2016) |
Expense ratios above are from the Vanguard VTI fund page. Note the one column where the stock wins outright: it charges you nothing. That is a genuine advantage, and it is worth roughly $3 a year per $10,000. Weigh it accordingly.
The Evidence Most Stock Pickers Have Never Seen
The strongest argument for the ETF is not a slogan about diversification. It is a dataset. In “Do Stocks Outperform Treasury Bills?” (Journal of Financial Economics, 2018), Arizona State’s Hendrik Bessembinder examined every US common stock in the CRSP database going back to 1926 — roughly 25,300 companies — and found:
- Four of every seven stocks (about 57%) had a lifetime buy-and-hold return lower than one-month Treasury bills. The median stock was a worse investment than cash.
- The best-performing 4% of companies account for the entire net wealth creation of the US stock market — nearly $35 trillion through 2016. Everything else, collectively, merely matched T-bills.
- The top 90 companies — about one-third of 1% of all stocks ever listed — produced more than half of all shareholder wealth ever created.
Read that carefully, because it reframes the whole question. Stock returns are not a bell curve where you are equally likely to land above or below average. They are wildly positively skewed: a small number of enormous winners drag the mean up while the typical stock quietly underperforms cash. An index ETF owns the winners automatically, by construction. A stock picker has to find them — and a portfolio of five or ten names is statistically likely to miss them entirely.
But the Professionals Can Do It, Right?
Mostly, no. S&P Dow Jones Indices publishes the SPIVA scorecard, which has tracked active managers against their benchmarks for two decades. As of the year-end 2024 US scorecard, 89.5% of actively managed US large-cap funds underperformed the S&P 500 over the trailing 15 years. These are full-time professionals with analyst teams, direct management access, and every data terminal money can buy, and roughly nine in ten of them lose to a fund that does nothing but hold the index.
That is not a knock on their intelligence. It is arithmetic: active investors collectively are the market, so before costs they earn the market return, and after costs they earn less. The index fund’s 0.03% is simply a smaller haircut than a stock picker’s research budget, spread, and mistakes. It is the same fee logic we quantify in our ETF fee drag tables, just applied to human effort instead of expense ratios.
Where Individual Stocks Genuinely Win
An honest comparison has to concede the other side, and there are four places a single stock beats a fund:
- No expense ratio. You pay the fund nothing because there is no fund. Small, but permanent.
- Lot-level tax-loss harvesting. You can sell one loser to offset gains while holding everything else. Inside an ETF, a losing position is invisible — you can only harvest if the entire fund is underwater. This is a real, recurring edge in a taxable account, and it interacts with the rules in our guide to how ETFs are taxed.
- Unlimited upside. A broad ETF is mathematically capped at the market’s return. A stock is not. If you are right about a company early, no fund will match it.
- Exact control. You decide precisely what you own — no unwanted sectors, no companies you object to, no index committee overruling you.
None of these overturn the base rate. They just explain why a thoughtful investor might want a slice of the portfolio in stocks rather than none.
The Structure That Uses Both: Core and Satellite
The resolution most experienced investors converge on is not to choose. It is to build a core from broad ETFs — the money that simply has to work — and to allow a capped satellite of individual stocks, commonly 5–10% of the portfolio, for genuine convictions. A 90/10 split means a stock that goes to zero costs you 10% of your portfolio rather than all of it, while a stock that ten-baggers still moves your total return meaningfully.
The discipline that makes this work is setting the cap before you start and rebalancing back to it, because the natural drift is for a winning satellite to quietly become the core. And the core itself should stay simple — one or two funds is usually the whole answer, as we work through in how many ETFs you should own. If you want the classic head-to-head on which broad fund makes the best core, our VOO vs VTI vs SPY comparison covers it. And since a satellite that works eventually creates a tax bill, our sister site’s guide to how capital gains tax is calculated is worth reading before you sell.
The Honest Summary
If you want the market’s return with near-zero effort and no chance of a single company ruining you, buy a broad ETF and stop. If you want the chance to do dramatically better and accept that the base rate says you probably will not, buy individual stocks — but size them like the lottery-shaped bets the data says they are. The mistake is not owning stocks. The mistake is owning five of them and calling it a portfolio.
Caveats
Expense ratios and holdings counts are current as of July 2026 from issuer fund pages and change over time; verify before buying. The Bessembinder and SPIVA findings are historical and describe base rates, not guarantees about any particular company or fund. Nothing here is investment advice, and past performance does not predict future results.