Status quo bias

The phone plan

Someone has paid $70 a month for the same phone plan for six years. Their carrier now offers a plan with the same coverage and data for $45, and switching takes a few minutes online. Every few months they notice the offer and think, “Mine works fine. I’ll leave it.”

The two plans differ only in price, and the cost of switching is a few minutes against $300 a year. Asked to choose between them as a new customer, they’d take the cheaper plan without a second thought. Staying put is what gives the old plan its advantage.

The inherited portfolio

A woman inherits her uncle’s investments, almost all in the shares of one regional bank. She would never have built a portfolio like that, and moving the money into a diversified fund would cost little. “It’s how he left it. I don’t feel right changing it.”

This is close to a scenario Samuelson and Zeckhauser used in their original experiments. It looks like caution or respect, which is why it’s less obvious than the phone plan. But the uncle chose the portfolio for his circumstances, and she’s choosing for hers. Whether she keeps it should turn on whether it suits her, and “it’s what’s there” doesn’t bear on that.

Enrolled by default

A company changes its retirement plan so that new employees are enrolled automatically at a 3% contribution, invested in a low-risk money market fund, unless they choose otherwise. Years later, most of those employees are still at exactly 3% in that fund, while employees hired before the change, who had to choose for themselves, mostly contribute more and hold mostly stocks.

This mirrors a real case studied by Madrian and Shea (see Evidence). Part of what keeps people at the default may be status quo bias: the preset option becomes the position they measure changes from. But the default may also be read as the company’s advice, and changing it takes effort. Those are different mechanisms, and the example shows why a default effect isn’t automatically status quo bias.

Samuelson and Zeckhauser described several conditions under which sticking with the status quo is rational, and research on defaults adds one more:

• When switching has real costs. Time, fees, learning a new system, or disruption can outweigh a modest improvement.
• When the alternative is more uncertain. You know how the current option performs; a new one might turn out worse than it looks.
• When the status quo carries information. Someone may have chosen it for reasons you can’t see, or you may once have analyzed the choice carefully and not remember why. See Chesterton’s fence.
• When the decision isn’t worth re-analyzing. For a small choice, keeping what works can be cheaper than working out the best option again.
• When the default is good advice. A preset option from someone who knows more than you and shares your interests can be a sensible thing to follow.

The test: choosing from scratch today, with the real costs of switching counted, which would you pick?

Staying with the family doctor

A new clinic opens nearby with slightly better online reviews than the family’s doctor of twelve years. The parents decide to stay put.

It’s a preference for the current option, but the doctor knows the family’s history, the records are there, and a new clinic’s reviews say little about how it would handle this family. Those are real switching costs and the uncertainty of the alternative, and they can justify staying.

The meeting nobody understands

A new office manager wants to cancel a Friday check-in call that seems pointless. A colleague suggests finding out why it was started first. It turns out to be when the overseas team, which has no other overlap, raises problems.

Keeping the call until its purpose is known is a preference for the status quo, but it rests on the status quo carrying information: an arrangement that has lasted may be doing something no one currently sees. Once the reason is known, the decision can be made on its merits.

Samuelson and Zeckhauser (1988) grouped explanations into three kinds. The first is rational: switching costs and uncertainty (above). The second is cognitive misperception, especially Loss aversion: measured from the status quo, what you’d give up by switching looms larger than what you’d gain. The third is psychological commitment: misperceived sunk costs (see Sunk cost fallacy), wanting to avoid the regret of a change that goes badly, and a drive to be consistent. Because they found the bias even in problems with no gains or losses to frame, they concluded it is “consistent with, but not solely prompted by, loss aversion.” That’s why this site doesn’t list it as a kind of loss aversion.

David Gal (2006) went further and argued that plain inertia explains status quo bias without loss aversion. In one of his experiments, over 85% of 110 students imagining they owned a quarter minted in one city said they would keep it rather than swap it for an identical quarter minted in another city, a choice with nothing to lose or gain.

Status quo bias and default effects are not the same thing. A default is what happens if you don’t choose. Jachimowicz and colleagues describe three ways defaults work: endorsement (the default is taken as a recommendation), ease (keeping it takes no effort) and endowment (the default feels like what you already have). Only the third corresponds to status quo bias as defined here. Status quo bias also appears without any default, as when an option is simply described as the current one.

The Endowment effect is a close relative: Kahneman, Knetsch and Thaler used a wine collector who would neither buy nor sell a bottle at the market price to illustrate both. Status quo bias is also different from an Appeal to tradition. That fallacy is an argument that something is right because it’s long established; status quo bias is a tendency in choice that needs no argument, and the status quo it favors can be days old.

• Run the from-scratch test. Ask which option you’d choose if you were deciding for the first time, then add the real costs of switching. If the answer changes only because one option is already in place, that’s the bias.
• Make the real costs explicit. Estimating the time, money and risk of changing separates legitimate reasons to stay from the pull of familiarity.
• When designing a choice for others, pick defaults deliberately. The field evidence shows that many people end up with whatever the default is, so a default is itself a decision.

These suggestions follow from the definition and the studies below, but haven’t been tested here as remedies.

Status: contested. The core pattern has survived a preregistered replication, and default effects in the field are large. But direct evidence for status quo bias itself, as distinct from defaults, comes mainly from hypothetical scenarios, and the replication failed in the one scenario where switching cost essentially nothing. Field default effects mix status quo bias with endorsement and effort, and a re-analysis correcting for publication bias found the evidence for the broader class of interventions that includes defaults undecided. Where the evidence falls between two labels, the standard is to choose the less confident one.

The foundational work. Samuelson and Zeckhauser (1988) gave 486 students hypothetical decisions, such as how to invest an inheritance, how to split a safety budget, or which color of car to take. Each problem came in a neutral version and in versions where one option was the status quo. For most options, the share choosing it was highest when it was the status quo, lower in the neutral version, and lowest when another option was the status quo. In field data, few Harvard employees switched health plans each year, and a 1986 survey found that only 28% of surveyed participants in the TIAA-CREF retirement system had ever changed how their contributions were split between its two funds, although almost all knew they could change it each year at no cost.

The replication. Xiao, Lam, Piara and Feldman (2021) ran two preregistered close replications of four scenarios with U.S. online participants (311 and 316 people). They found strong support for status quo bias in three: the budget allocation, the investment portfolios and a choice between college jobs, with effects comparable to or larger than the originals. They found no substantial support in the fourth, a choice of color for a car one had ordered. They noted that switching color cost essentially nothing and the status quo there was simply the color the dealer said was available, while in the other scenarios participants may have inferred costs of changing, and they recommended “caution regarding the claim that the bias is pervasive.” They also pointed out that they tested options described as the status quo, not people’s actual past choices.

Field evidence on defaults.

• Madrian and Shea (2001) studied a large U.S. company that switched to automatic enrollment in its 401(k) retirement plan. At 3 to 15 months of tenure, 37% of employees hired shortly before the change were participating, against 86% of those hired under automatic enrollment. About three-quarters of automatically enrolled participants stayed at the 3% default contribution rate, and most of their money went into the default money market fund, choices few earlier employees had made. The authors attributed this to inertia and to employees taking the default as advice.
• Johnson and Goldstein (2003) compared European countries where people are organ donors unless they opt out with countries where they must opt in. Effective consent rates ranged from about 4% to 28% in the opt-in countries and from 86% to nearly 100% in the opt-out countries. In an online experiment with 161 people, 42% agreed to be donors when that required opting in, 82% when it required opting out, and 79% with no default. Actual donations differed much less: in their analysis, presumed consent was associated with a 16% higher donation rate.
• Jachimowicz, Duncan, Weber and Johnson (2019) meta-analyzed 58 default studies with 73,675 people and found a sizable average effect (d = 0.68), with wide variation, including some null and two negative results. A trim-and-fill correction for missing studies raised the estimate (d = 0.80), though the authors cautioned that such tests are unreliable when results vary this much. Defaults worked better when they signaled endorsement or felt like the status quo; ease of keeping the default was not a significant moderator.

Doubts about the size of default effects. Mertens and colleagues (2022) estimated d = 0.62 for defaults in a meta-analysis of choice architecture (“nudge”) interventions. Re-analyzing those data with a method that corrects for publication bias, Maier and colleagues (2022) found no evidence left for nudges overall, and undecided evidence for the category that contains defaults. They noted that heterogeneity means some nudges may still work.

What remains uncertain. People often stay with whatever is presented as the current option, and in real settings defaults can shape outcomes for years. What’s unsettled is how much of that is a bias rather than a response to real costs, uncertainty or implied advice; how pervasive it is across kinds of decisions; and whether it is explained by loss aversion, inertia or commitment.