Gains and Losses: The Arbitrary Effect of Decision Frame

The outcome of a lawsuit can sometimes turn on an arbitrary reference point.

|

Imagine that you bet a friend $50 that the Kansas City Chiefs would win Super Bowl LV in February of 2021. In a triumph of experience over youth, Tom Brady and the Tampa Bay Buccaneers beat Patrick Mahomes and the Chiefs. How would you describe the outcome of your bet? Most people would say something like, "I lost $50." That is only partly accurate, however. You have not only lost $50, but you have also failed to win the $50 that your friend would have paid you had the Chiefs won. In reality, Tom Brady's outstanding performance cost you $100 (sorry Brady-haters, it was outstanding).

The bet above demonstrates the widely studied, deeply powerful phenomenon that psychologists Amos Tversky and Daniel Kahneman called "framing." Framing is the tendency to treat an improvement from the status quo (gains) differently than a deterioration from the status quo (losses). Losses hurt more than identical gains feel good. In the Superbowl bet, the loss of $50 attracts much more attention than the foregone gain of $50. People work harder and take more risks to avoid losses than to obtain comparable gains. They also value their possessions more than comparable goods that they do not yet own (the endowment effect). Treating gains and losses differently can be sensible, but it often occurs when the status quo is arbitrary or even meaningless.

The influence of framing can be found in many areas of law. The Supreme Court interprets constitutionally enshrined federalism as precluding the federal government from imposing penalties (losses) on states for failing to enact legislation, but federal statutes that withhold federal highway funds (foregone gains) for failing to enact legislation are constitutionally acceptable. A New York statute forbids the use of surcharges (losses for those who use credit) on credit card transactions but allows discounts for cash (foregone gains for those who use credit). The Clean Air Act regulates new sources of air pollution (which can reduce potential profits from future facilities) far more aggressively than existing sources (which would impose losses on existing facilities). As the late Professor Patrick Atiyah put it: "To deprive somebody of something which he merely expects to receive is a less serious wrong, deserving less protection, than to deprive somebody of the expectation of continuing to hold something which he already possesses."

In research on sitting trial judges, we have found that framing has a big influence on how judges think. We conducted a series of eight experiments with trial judges. In each experiment, judges read a hypothetical case in which we varied the reference point of a lawsuit; half of the judges read a case involving a foregone gain for a litigant, while the other half read the same case recast as involving a loss for that litigant. For example, in one study, we asked trial judges to award compensatory damages for medical malpractice that left a patient permanently blind. For half of the judges, the surgery was necessary to restore vision to a patient who had recently lost it; for the other half, the surgery was necessary for the patient to avoid losing his vision. The judges awarded a median of $1.5 million in compensation for an operation that failed to restore the plaintiff's eyesight, but $2.5 million for an operation that deprived the plaintiff of his eyesight. The judges thus assigned a higher price to the loss of sight than to the foregone opportunity to regain vision.

In another study, we explored whether judges would react differently to evidence of age discrimination based on whether the alleged discrimination occurred in firing an employee as opposed to failing to hire a job applicant. We informed the judges of a state statute forbidding an employer from discriminating on the basis of age—a prohibition that applied with equal force to both hiring and firing. For half of the judges, the complainant was one of five finalists for a residence director job at a local college who ultimately lost out to four much younger applicants. For the other half, the complainant was one of five residence directors faced with a layoff due to staff reductions, and whom the college ultimately let go instead of four much younger employees. In both cases, the evidence that the college had favored younger applicants or employees was identical. Only 14% of the judges determined that age was a "substantial motivating factor" in the college's decision not to hire the complainant, compared to 30% of the judges who evaluated the decision to lay off the complainant. The same evidence seemed more compelling when the employee had been terminated than simply not hired.

In a third study, we asked judges whether they would rescind a contract based on mutual mistake. For half of the judges, the plaintiff seeking rescission was a buyer who had purchased a collectible he had thought was valuable but which actually turned out to be worthless. For the other half, the plaintiff was a seller who had sold a collectible he had thought was worthless but which was actually valuable. In both cases, the parties were ignorant of the item's actual value and had made no representations about its worth. Upholding the contract imposed a loss on the disappointed buyer but was a foregone gain for the disappointed seller. Although the doctrine of mutual mistake does not distinguish between buyers and sellers, the judges in our study did. Among those evaluating buyers, 82% favored rescission, as compared to only 41% of those evaluating sellers. We also found the same framing effect among a group of elite arbitrators who considered this same problem.

We also found that framing had a robust influence on contractual damages, wage disputes, water rights, settlement offers, liability for vaccines, and bankruptcy reorganization plans. In all of our experiments, the frames created an arbitrary influence. A lifetime of eyesight is not more valuable when it is lost rather than when a simple surgery that should restore it fails. Likewise, being fired is usually worse than not being hired, but the same evidence should not be evaluated as more probative of discrimination merely because firing rather than hiring is at stake. And finally, the mistake in our contracts case was just as costly to the seller as it was to the buyer, even though judges treated them differently.

An arbitrary reference point creates a powerful illusion. So long as they recognize the influence that frame has, however, judges can "imagine the opposite" by considering whether the case would come out differently if the stakeholder were in the opposite position. Also, even though research shows that lawyers are at least as vulnerable to the influence of framing as judges, lawyers can try to reframe how they present cases to judges. We suspect, however, that even with such efforts, frame will continue to influence case outcomes.

Our full article on framing in judges: Jeffrey J. Rachlinski & Andrew J. Wistrich, Gains, Losses & Judging: Framing the Judiciary, 94 Notre Dame L. Rev. 521 (2018).

Tomorrow: Emotional Influences on Judges

NEXT: Today in Supreme Court History: June 10, 1916

Editor's Note: We invite comments and request that they be civil and on-topic. We do not moderate or assume any responsibility for comments, which are owned by the readers who post them. Comments do not represent the views of Reason.com or Reason Foundation. We reserve the right to delete any comment for any reason at any time. Report abuses.

  1. In general, you have dismissed the very practical nature of this difference.

    Money that you might have gained in a bet is money of little value because it’s uncertain. It was never yours (and you can never be sure the guy was going to pay it). However, money you lost was hard cash that you worked for. You gave hours of your life for it.

    On the other hand, there is a very practical reason environmental regulations regulate existing sources lighter than new ones. A 40 year old engine might have been the best of its time, but we can make much better ones now. However, forcing companies to continuously spend tens of millions to replace equipment early is both polluting (because making engine upgrades takes emissions itself) and likely to send all the jobs overseas due to the waste.

    On the other hand, requiring a new engine to meet a stricter NOx requirements has relatively low costs since the only difference is the upgrades from what the system would have cost without it, rather than replacing the entire system.

    1. Good explanation of the environmental framing question.

  2. I’ve been saying for years government letting you keep your own money you earned is not subsidizing you. That’s a rhetorical technique to make it seem like you are getting an unfair benefit, so feelings of jealousy can rise, ultimately to benefit some politicians.

    Note this is not a defense of such policies, call them loopholes, or “economic policies”, on principle or in practice.

    1. The reason is obvious. It’s like saying a thief who declines to steal is giving you stuff. It’s important not to philosophically disable yourself to become more akin to an abuse victim no longer resisting due to wool pulled over their eyes.

    2. Interest, but also a case of framing. You could say the government is letting you keep your money or you could say the government is not asking you to pay for services your are getting. What the government is often doing by not taxing enough is to deferring payment by borrowing money. So you are now paying for the services as well as interest for deferring timely payment.

  3. Judges should be replaced with an algorithm written and owned by the legislature. There should be liability for any error of programming or of execution. Machines are 100 times better. Compare a commute to work on a horse to one in a car on a snow day.

  4. Most of the examples here of reference point bias are good ones, but I have to disagree with the one about employment. The “moral arithmetic error” is numerically equating hiring with firing when the correct equivalence is hiring with not firing.

    If you are going to employ one person out of pool of four (by hiring) versus employ three people out of a pool of four (by not firing), it’s clear that even with no bias the criteria/standard for employment in the first case necessarily *must* be tighter. Or if you prefer to view it negatively as discrimination, in the first case one *must* discriminate against three people on some basis, valid or not, while in the second case you must discriminate against only one. Again, the basis must necessarily be different.

  5. I’ve always considered the ACA mandate/penalty/tax issue an example of this phenomenon.

    We have all kinds of tax credits for doing things, but when we institute a penalty for not doing something we get a lot of screaming about broccoli,

    1. A couple problems here. First, you’re engaging in the Starving Child of Somalia fallacy.

      One person fails to send a $100 donation each month to Somalia, and a child there dies of starvation and neglect who could have been saved. Another person flies to Sweden, kidnaps a healthy child, and keeps him locked in a basement until he’s dead. The fallacy is saying the two situations are the same because the net result is one dead child.

      More generally the fallacy is failure to recognize that action and inaction are fundamentally different categories. We can only do a finite number of things in life, but we can refrain from doing an infinite number of things. It’s also why compulsion (e,g. be a military conscript for three years) is inherently more intrusive than a superficially equivalent prohibition (e.g. you are forbidden to join the military for three years).

      The second problem is your failure to recognize that pretty much all decent societies believe that even very minor *punishment* requires some kind of individualized due process and a finding of wrongdoing. We are much more careful about fining someone $100 for speeding than hitting them with $2000 in property tax. People seek to have wrongful convictions reversed even after they’ve finished their sentences and can never get the time back.
      Why? It’s about justice and opprobrium rather than dollars and years.

      The ACA proponents shot themselves in the foot by calling it a *penalty* because they just couldn’t resist inflicting some opprobrium on the uninsured. They wanted to make a collectivist moral statement about obligations to society and were willing to risk the constitutionality of their program to do so. If they had just quietly raised taxes a few hundred dollars in one bill and then offered the credit in another, this wouldn’t have happened. But they had to have their little slap at individualists, and then act astonished when people object to being slapped.

      1. The ACA proponents shot themselves in the foot by calling it a *penalty* because they just couldn’t resist inflicting some opprobrium on the uninsured. They wanted to make a collectivist moral statement about obligations to society and were willing to risk the constitutionality of their program to do so. If they had just quietly raised taxes a few hundred dollars in one bill and then offered the credit in another, this wouldn’t have happened.

        Which is pretty much my point. The difference is semantic only, and that’s what got opponents upset or, less generously, gave the opponents a cudgel.

        Nor is there “coercion” in any meaningful sense. The penalty did not “coerce” you to get insurance. It gave you a choice between getting it or paying the penalty. It’s unlike a speeding ticket because you can get any number of speeding tickets a year, but you’ll pay the penalty only once.

        1. Calling someone a wrongdoer is not merely a “semantic” issue. It’s something one can object to even if there are no financial consequences attached. And by the way, I do object.

          True story: one time I overpaid my taxes, by counting sale of a patent as regular income instead of capital gains. Here’s how the IRS chose to frame it: they rejected my *entire payment* because it was the wrong amount, but held back the correct taxes plus a fine. The fine was recorded for “failure to pay taxes”.

          I did not disagree with the overall financial outcome, however, I thought the fine should have been recorded for “overpayment of taxes”. The IRS took the reductive position Bernardist posting that it doesn’t matter, since the dollars come out the same.

          They didn’t seem to understand that their is a moral difference between evading taxes and overpaying taxes, and that it can make a real difference. For example, if one was applying for some position on a charity board or a minor political office, having evaded is a disqualification and having overpayed is probably not.

    2. Let me put in another way, Bernard.

      Suppose an abortion costs $900. Some state decides that they will reimburse doctors who perform one $900. To fund the program, they will criminalize women who have get an abortion. It will be a Class C Misdemeanor with a fine of $900.

      I assume you see why there would be an objection.

      1. I don’t get your point here.

        1. The Bernardist position, as I understand it, is that either way the abortion costs the woman $900, so it’s just a “semantic” distinction.

          My position is that the woman rightly objects to being labelled a wrongdoer. Sometimes it really isn’t about the money.

          1. The point being that I object to being labelled a wrongdoer if I decided to take the “no insurance” option and that is an independent consideration from the net number of dollars it costs me.

          2. Well, yes, it is just a semantic distinction.

            Would anyone argue that a company that didn’t do whatever R&D is required for the R&D tax credit was doing something wrong?

            Regardless, the ACA created no misdemeanor, did not label anyone a “wrongdoer.” Unlike other circumstances there are no secondary consequences. If I pay a penalty because I file my taxes late I’m not labeled some kind of bad guy. The only embarrassment it might cause if I tell others has to with my own negligence, not some harsh law.

            OK. I’ll give you that you interpret things that way, and don’t like it, and I can understand it, even though I don’t think it’s wholly rational. But the fact that someone might take offense at having to pay a “penalty” rather than a “tax” is a vanishingly slim reed on which to base a claim of unconstitutionality.

            In any case, you continue to reinforce the point I started with – that this is a prominent example of how the framing of an issue affects people’s attitudes.

  6. Losses hurt more than identical gains feel good. In the Superbowl bet, the loss of $50 attracts much more attention than the foregone gain of $50. People work harder and take more risks to avoid losses than to obtain comparable gains.

    Just to nitpick a bit, this is really what Kahneman and Tversky called “loss aversion,” partly to distinguish it from “risk aversion.”

    “Framing” generally describes influencing a decision by the way equivalent alternatives are described. K and T give an example:

    Imagine a disease that is expected to kill 600 people, but there are some ways to fight this.

    In a study, one group is asked to choose between two alternatives:

    Under Program A 200 lives will be saved.
    Under Program B there is a 1/3 chance that 600 lives will be saved, and a 2/3 chance that no one will be saved.

    A second group is given the following choices:

    Under Program C 400 people will die.
    Under Program D there is a 1/3 chance that 600 lives will be saved, and a 2/3 chance that no one will be saved.

    In the first group the majority chooses A, in the second the majority chooses D, even though A and B are exactly the same as C and D.

    When the issue is framed as a matter of loss, loss aversion kicks in, and the majority chooses the riskier scheme.

  7. I remember, long long ago, in a galaxy far away, I invented (with help) a packaged service that I thought several large multinational companies might like to buy – that being my job at the time. I tested it on one of the Board members of my employer, who had had lots of big executive jobs over the years.

    His reaction was favorable – “sounds excellent ! How much ?”
    “Er, I though about $250,000”
    “Oh ! Oh no ! You offer me all that and you’re only asking $250,000 ? You should ask for at least $2 million. Asking for only $250,000 makes it seem, well….only worth $250,000 !”

  8. “In reality, Tom Brady’s outstanding performance cost you $100”. Um, no. In *reality* it cost you $50 and you lost $50. *Hypothetically*, you are down $100 from your best possible – but uncertain – outcome. If you had WON the bet, but your friend refused to pay you, then you are down $100. Your system doesn’t distinguish between the two, which makes it incomplete and misleading.

    1. A place I worked had an accountant who was in the habit of counting financial chickens before they had hatched. Also had a colleague who was always spending next month’s paycheck this month. The accountant ended up getting fired and the colleague ended getting his bank account frozen.

  9. While I am not a Brady fan I have to say his performance is assume. In the game against my beloved Packers Brady threw 3 interceptions and yet he continued to throw the ball. Very few QBs are mentally strong enough to put three interception behind them and continue to function. He is just better than all the rest.

  10. I think framing arguments like this are highly overrated in the law, because they ignore the fact that if judges or juries see through your framing, you look less honest.

  11. I think there’s an obvious rationale for treating foregone gains and losses differently. Losses involve a certainty. But gains only involve an expectation, a probability. You expected to gain something had an intervening event not occurred. There’s a huge difference between a fact and a counterfactual. Huge beings are not prophets. We see only backwards, not forewards. Facts happened. But you have to make assumptions to make statements about what could or should or might or ought to have happened, or what will happen. There’s a reason for the old saying that a bird in the hand is worth two in the bush. They don’t have the same value at all.

    So in nearly all the examples given, the counterfactual has to be ddiscounted by some uncertainty of its occurring, while the fact doesn’t. We know for sure that the fired employee met the company’s hiring criteria. He was hired. But would the not-hired employee have been hired but for the doscriminatory reason? We don’t know for sure. The same is true for the other examples as well. Will the person in danger of losing his sight actually do so? The blind patient has for sure, but the other patient might not, or might not for a long time. And so on.

    Classical economics assumes a certainty about the world, and aboit the future, that human beings simply do not posess. When Tversky and Kahneman’s work first came out, it was said that their work showed that hunan beings don’t behave rationally. Not so. More often, it shows that classical economics and its rational choice theory is based on overly naive and overly optimistic assumptions about the things like the knowledge humans possess when they make choices.

  12. Imagine that you bet a friend $50 that the Kansas City Chiefs would win Super Bowl LV in February of 2021. In a triumph of experience over youth, Tom Brady and the Tampa Bay Buccaneers beat Patrick Mahomes and the Chiefs. How would you describe the outcome of your bet? Most people would say something like, “I lost $50.” That is only partly accurate, however. You have not only lost $50, but you have also failed to win the $50 that your friend would have paid you had the Chiefs won. In reality, Tom Brady’s outstanding performance cost you $100

    This is… nonsense. In order to claim that you lost $100 when the Buccaneers defeated the Chiefs, you would have to also claim that — if the Chiefs had defeated the Buccaneers — you would have gained $0. There is no frame of reference in which that makes any sense.

    Look at the amount of money you have over time, and you’ll see it drop by $50 when the Chiefs lose, or increase by $50 when the Chiefs win. But recognizing that that’s happened is a mistake… in only one of those cases?

    1. I think you’re right.

      Did I lose millions of dollars because a stock I bought didn’t go to the moon?

      There has to be a baseline against which to measure gains and losses. Here that’s zero. You either win 50 or lose 50.

      Consider, again, the stock market. How much do you lose when your portfolio goes down $100? Certainly not the gain you would have had had you made the most profitable possible investment. If that were so, a gain of $100 would be a loss if another portfolio would have gained $200.

      You could argue, I suppose, that you lost money if you made less than say, the S&P500, or that you would have made in a risk-free investment. That’s a useful metric in some circumstances, but even there saying you lost money is a long stretch.

    2. It’s not nonsense – but it is an attempt to say in English what is better said in math. If you start with $x and the Chiefs win (scenario A), you end with $x+50 but if the Buccaneers win (B), you end with $x-50. The difference between A and B is ($x+50)-($x-50) = $100.

      So the difference in outcome is $100. That’s what he meant when he said that Tom Brady’s performance cost you $100. The mistake is in thinking that it only cost you $50. There is no scenario C (in the hypothetical presented) where you choose not to bet – yet that’s the baseline that we naturally think from even when the bet’s already been make and C is not even theoretically possible anymore.

      1. Aargh!!! Can we please have an edit button?
        “been make” should have been “been made”

      2. How do you deal with odds? If I lose on a $1 lottery ticket with 1,000,000 to 1 payout, am I 10 times worse off than if the payout had only been 100,000 to 1? I don’t see how this makes any sense.

        1. Odds don’t matter in a statement of outcomes. Yes, the difference in outcome is about ten times larger in your first wager than in your second.

          Now, if you’re trying to decide whether to take the bet, the odds do matter. The conventional understanding is that breakeven between the two bets (assuming you must choose one) is when the odds for the smaller bet are 10 times the odds of winning the larger. So if you have a one-in-a-million chance of winning the larger bet, you need a one-in-one-hundred-thousand chance for the smaller.

          The implication of the concept above, though, is that the odds that justify the smaller bet aren’t exactly 10 times the odds that justify the larger bet – they are 999999/99999. That is, you need odds that are 10.00009 better for an equal probability-adjusted outcome.

          1. But that’s the problem right there. odds DO matter when making decidions.

  13. I’m not sure this is due to the frame. It may have to do with accounting for probability (even if that’s not officially permitted.)

    For half of the judges, the plaintiff seeking rescission was a buyer who had purchased a collectible he had thought was valuable but which actually turned out to be worthless. For the other half, the plaintiff was a seller who had sold a collectible he had thought was worthless but which was actually valuable.

    I’d do the same. The reason is that the seller had the item previously and likely had ample opportunity to verify its value. A buyer often has less opportunity to check. Beyond that, there is always a possibility of some misrepresentation by the seller which may not be easy to prove.

  14. Situations are different, and a lot depends on the qualifications and professionalism of the lawyer. An indicator of any activity is a result, and it can be either positive or negative. In the case of interaction with a professional lawyer https://www.lawyersuae.com/ , the result is checked by the fact that your problem is solved or it did not occur at all if you asked for help in advance.

Please to post comments