Ostrich Effect: Why People Look Away From Bad News

Information has two jobs, and they pull in opposite directions. It tells you what to do, and it also tells you how to feel. Karlsson, Loewenstein and Seppi's account starts from the second job: paying attention to your portfolio amplifies whatever the news is, so attention is worth having when the news is likely to be good and worth avoiding when it is not. On that account the asymmetry is not a failure of nerve, it is a trade, and the price is whatever you give up by not knowing. The name comes from a myth, since ostriches do not bury their heads in sand.

The effect is visible because account log-ins are recorded. Sicherman, Loewenstein, Seppi and Utkus measured daily log-ins for 100,000 accounts over 2007 and 2008 and found that the total fell by an estimated 9.5 percent the day after a market decline, with the same pattern over the previous week and the previous month, and a further drop when the VIX volatility index was high. Attention also rose with the amount of stock market coverage in the news, so the avoidance is specific to personal numbers rather than to the subject.

Two findings in that paper are worth more than the headline. Avoidance behaves like a trait: the same investors showed it across daily, weekly and monthly horizons and across both years. And the pattern disappears where the theory says it should. Among account holders who held no equities, for whom a market move says nothing about their own balance, the ratio of ostriches to the opposite type flipped, from about four to one in the main sample to roughly one to two.

Two limits keep the finding honest. The relationship between returns and attention is not a straight line: after large declines, log-ins rise again, which the authors read as people checking whether to act, so the ostrich pattern describes ordinary bad weeks rather than crises. And it is not universal. Gherzi, Egan, Stewart, Haisley and Ayton analyzed the log-in records of a smaller sample of active online investors, alongside personality survey answers, and found the opposite pattern, which they called the meerkat effect: monitoring went up after daily market gains and after daily losses, and differed by personality. Arna Olafsson and Michaela Pagel found the financial version of the same asymmetry outside the market, in a banking app: attention fell as spending and overdrafts rose, and rose with cash on hand and on payday.

Niklas Karlsson, George Loewenstein and Duane Seppi published "The ostrich effect: Selective attention to information" in the Journal of Risk and Uncertainty in 2009. The term had been used three years earlier by Dan Galai and Orly Sade for a different claim about financial assets, and this paper gave it the meaning it now has. The authors proposed that attention amplifies the hedonic impact of information, which predicts that investors will look at their portfolios more often after the market has risen than after it has fallen, even though the information is equally available either way and equally useful.

Their test used aggregate data rather than individuals: the daily count of account look-ups at an American investment company, Vanguard, and at the Swedish Premium Pension Authority, set against weekly market index returns. Look-ups followed the market up and dropped when it fell. Because the data were aggregated, the paper could not say whether the same individuals were avoiding information or whether different people were logging in, which is the gap the next study closed.

Nachum Sicherman, George Loewenstein, Duane Seppi and Stephen Utkus published "Financial attention" in the Review of Financial Studies in 2016, using daily log-in records for 100,000 paperless Vanguard accounts across 2007 and 2008, part of a full sample of 1.2 million accounts. Log-ins fell by an estimated 9.5 percent the day after a market decline and fell with high market volatility, and the authors could now classify individual investors. Men, older investors and wealthier investors paid more attention in general and were also more likely to show the avoidance; the classification was stable over two years; and investors who avoided information were less likely to trade after downturns.

The 2007 and 2008 window is both the strength and the weakness of that study. It covers an extraordinary period, which is how the authors were able to see that the effect reverses after very large falls, and it means the size of the effect in calmer years is a separate question. Both papers measure attention to numbers, not emotion, and neither can say what a given investor was feeling when the statement went unopened.

Any number you can choose not to look at will do. The rows below are illustrations, not findings from the studies cited on this page.

The useful question is not whether you look often but whether you look evenly, so the counters all remove the decision from the moment the news arrives.

• Put the check on a schedule. A quarterly date in the calendar is the same on a bad week as on a good one, which is the whole point.
• Automate what the information is for. Automatic contributions and automatic rebalancing mean fewer weeks in which not looking has a cost.
• Separate looking from acting. Much avoidance is really avoidance of a decision, and agreeing with yourself that you will only read the number makes it easier to open.
• Have the numbers delivered. An alert, a dashboard emailed on Mondays or a colleague who presents the figures removes the step where you choose whether to see them.
• Notice the asymmetry rather than the frequency. If you knew the market was down without opening the statement, you had the bad news already; what you avoided was the part that applied to you.
• Accept that some avoidance is reasonable. For a portfolio you do not trade, a weekly figure carries no decision. The costly case is information you need in order to act.

Three neighbors concern the same traffic between a person and unwelcome information.

• It is not information bias, the appetite for information that cannot change a decision. That is the opposite behavior with the same underlying question: the ostrich effect skips information that would be useful, information bias collects information that would not.
• It is not normalcy bias, the assumption during a warning that things will carry on as usual. Normalcy bias is a belief about what is happening; the ostrich effect is a choice not to find out.
• It is not confirmation bias, the search for evidence that fits what you already believe. Confirmation bias picks which information to take in; the ostrich effect turns the tap off entirely when the news is likely to be bad.