The Stocks You Sell Beat the Stocks You Buy: Do Investors Trade Too Much?

Standard finance has an awkward relationship with trading volume. In a market of rational agents who all know each other are rational, there is very little reason to trade at all. If someone wants to sell me a stock, I should ask what they know that I don't, and decline. Rational traders should mostly sit still.

Real markets trade colossal volumes. Something is generating all this activity, and it is not information.

Terrance Odean asked the most direct possible question: are all these trades any good? Not "do investors beat the market", which is a question about skill in general. Something much sharper: when an investor sells stock A and buys stock B, does B subsequently beat A?

Because if it doesn't, then the entire trade was worse than useless. It was a lateral move that cost money.

The problem: measuring whether a trade was worth making

This is a beautifully clean test, and it is the paper's whole contribution.

Forget benchmarks. Forget alpha. Forget factor models. When an investor swaps one stock for another, they are making an implicit, specific forecast: the stock I am buying will do better than the stock I am selling. That is the only claim a switch trade makes, and it is directly checkable.

So Odean took the discount brokerage records (the same rich dataset behind his disposition effect work) and did exactly that. He identified the trades, then simply followed both stocks forward and compared. The stock they bought, versus the stock they sold, over the following year.

Note what this test elegantly avoids. It does not need a risk model. It does not need to argue about what the correct benchmark is. It simply asks whether the swap improved the portfolio, using the investor's own decision as the hypothesis.

The key idea via analogy: the restless gardener

Picture a gardener who is convinced he has a gift for plants. Every week he pulls up a plant that isn't thriving and replaces it with a new one he is certain will do better.

At the end of the year, we check on the plants he ripped out (they got replanted elsewhere) and compare them to the ones he chose. And the plants he tore out are, on average, doing better than the ones he put in.

Every switch made the garden slightly worse. And he did this every week, paying for new plants each time and trampling the soil in the process.

That is Odean's finding. The stocks investors sold subsequently outperformed the stocks they bought. And this is measured before counting the costs of trading. Once you add in the commissions on both the sale and the purchase, and the bid-ask spread paid on both, the trade is a substantial, reliable destruction of wealth.

Read the logic carefully, because it is stronger than a simple "investors underperform" result.

If investors simply picked stocks at random, then on average the bought stocks and the sold stocks would perform about the same, and trading would cost them only commissions. That would be bad enough. But Odean finds something worse: the swap itself has negative value even ignoring costs. Investors are not merely paying for nothing. They are paying to make their portfolios worse. Their stock-picking signal is not zero, it is inverted.

The natural explanation, and the one Odean gives, is overconfidence. People trade because they believe they know something. If your belief in your own information is well-calibrated, you trade only when you truly have an edge, and your trades add value. If you systematically overrate your private signals, as the psychology literature says people do, you will trade far more often than your genuine edge justifies, on signals that are mostly noise, and every trade costs you a spread and a commission.

He also finds that this is not driven by rational reasons to trade. Some trades happen for perfectly good reasons: rebalancing, liquidity needs, tax-loss harvesting, moving money out for a house deposit. Odean checks these. Those explanations account for some of the trading, but nowhere near enough. The bulk of trading is people switching horses because they think the new horse is faster, and it isn't.

Why it mattered

• It is the empirical test of overconfidence in the wild. Daniel, Hirshleifer and Subrahmanyam had built a whole theory of markets around investor overconfidence. Odean showed it directly: real investors trade as if their information is valuable, and it demonstrably is not.
• It links psychology to a hard, painful number. Not "investors are biased" but "investors' trades destroy value even before costs, and the costs then destroy more". That gets attention in a way that a laboratory result never does.
• It became the single best argument for passive investing. The case for index funds is usually made on the basis of fees. Odean makes a deeper case: even if trading were free, the average retail investor's decisions would still be value-destroying, because the decisions themselves are worse than random. Doing nothing is not just cheaper, it is a genuinely better strategy.
• It set up the follow-up that put a number on the damage. With Brad Barber, Odean would go on to show that the households who trade the most earn dramatically lower net returns than the market, and that men, who psychology says are more overconfident about finance, trade substantially more than women and pay for it.

The honest limitations

• The overconfidence interpretation is an inference. Odean sees excessive, value-destroying trading. He does not directly measure anyone's confidence. Overconfidence is the most natural explanation but it is imported from psychology, not proven in the data. Alternative stories (entertainment value, sensation-seeking, gambling preference) fit the same facts, and the later literature suggests they all contribute.
• It is retail, not professional. These are discount brokerage clients: self-directed, non-institutional. The result says nothing about whether professionals trade too much. It is tempting to extrapolate, and probably somewhat justified given how professional fund performance looks, but this paper does not show it.
• Averages hide the distribution. "On average, trades destroy value" does not mean every investor's trades destroy value. Some investors are genuinely skilled. The average is dragged down by a mass of overconfident traders, and the paper does not spend much time on the right tail.
• It does not tell you what to do instead, beyond "less". The prescription implied is: trade less, index more. Fine advice, and probably correct for most people, but it is a conclusion about the average investor being handed to individuals who all believe they are above average, which is precisely the bias the paper documents.

The one-line takeaway

Odean checked whether investors' trades actually helped, and found that the stocks they sold went on to beat the stocks they bought, before commissions, meaning the average investor is not just paying fees to break even, they are paying fees for the privilege of making their portfolio worse.