Psychologist Dan Ariely, who writes an “Ask Ariely” column in the weekend Wall Street Journal, leads in last weekend’s column with a psychological explanation of a phenomenon that often arises in auctions. In doing so, he fails to consider an obvious economic explanation.
Questioner Liron writes:
Recently I was at an auction of antiques that included some pretty unappealing items. For one item in particular, the auctioneer started the bidding at $20, but no one was willing to go that high. The auctioneer then reduced the price to $10, and at that price, many people began to bid. In fact, the bids quickly went above $20. How can it be that an item no one wanted to pay $20 for ended up selling for much more?
Ariely answers:
Your story is a good illustration of the idea of “auction fever.” In auctions, people have to make an initial decision whether or not to participate in the bidding, and the starting price plays an important role. In this case, $20 seemed like too much to pay to get into the auction.
But once we enter an auction and start bidding on an item, we begin to think of it as ours, which makes us like it more. Instead of asking ourselves “Do I want to pay this price for this item?” we start thinking, “I am not going to let someone take this item away from me.” The result is that we end up paying a higher price because of our feelings of attachment.
Do you see the obvious explanation Ariely omits? Let’s say a number of bidders know that they value the item at $20 or more. What they don’t know is that there are others in the room who also do. So when the auctioneer opens with $20, various people willing to pay more might think that, correctly, it turns out, if they don’t bid, the auctioneer will lower the opening ask. So the auctioneer cuts it to $10. Then someone who values it at more than $10 bids $10. And with that, the bids rise. As they progress, various bidders find out more about what other bidders’ minimum value is. So it’s easy to understand how multiple bidders go above $20.
Is Ariely’s explanation plausible? I guess. But what’s shocking is that Ariely doesn’t even consider a straightforward economic explanation.
In that same column, in response to the final questioner’s query, Ariely writes:
One key to understanding how politics works is the concept of “motivated reasoning”—our tendency to interpret data in a way that is consistent with our worldview, to see only what we want to see.
Do you see the irony?
P.S. For more on auctions, see Leslie R. Fine, “Auctions,” in The Concise Encyclopedia of Economics.
READER COMMENTS
Jon Murphy
Jun 17 2019 at 12:13pm
I like your explanation better; it’s more simple. Ariely is assuming a rather large thing: knowledge of the inner workings of the mind of the bidders (“But once we enter an auction and start bidding on an item, we begin to think of it as ours, which makes us like it more”). Both the impartial spectators here (Ariely and the correspondent) do not have that kind of information; it is merely assumed in. If that assumption does not hold, the explanation falls apart and we are in no better place than we were before.
Conversely, the economic answer is straightforward and doesn’t require as many assumptions. All we are assuming is people have an idea of what they value an object at and think of a way to get it at as low a price as possible. There’s no need to assume irrationality on their part (“auction madness”) or changing perspectives. It’s more robust without losing explanatory power.
David Henderson
Jun 17 2019 at 2:20pm
Very nicely stated. Actually you stated it better than I did.
Jon Murphy
Jun 17 2019 at 3:43pm
Thank you
David Seltzer
Jun 17 2019 at 5:09pm
Jon, Well stated. I was a market maker on the CBOE and American Stock exchange.
Seldom were there informational asymmetries in the equity markets. A trader had the same information other traders had. If asymmetries emerged, they were quickly arbitraged away. Impounded information rendered the markets efficient. This was reflected by the narrow spread between the bid and offer for both the options and the underlying securities.
john hare
Jun 17 2019 at 12:29pm
I buy our work trucks at auction most of the time. And a little group helps others save money. I never take the starting ask even if it’s within my limit for the item. Often they will start at $5k on a vehicle and work it down until someone bids $1.5k or so. If I really want it, I use a strategy that intimidates some bidders. If bids are at $1,500.00 and going in $100.00 increments, I’ll bid $2k. People that are trying to low ball balk at a $500.00 jump where they might creep up to $3k=$4k a hundred at a time. If someone then bids $2,100.00, I’ll go $2,500.00. And walk away if they keep going.
It’s not that I wouldn’t pay $5k if I needed to, it’s that I go to auctions to save money. A few months ago our group got a college student into a 2013 C-MAX for $1,500.00 using this strategy.
Jon Murphy
Jun 17 2019 at 12:40pm
I know who I’m going to the next time I need a vehicle 😉
Dylan
Jun 17 2019 at 12:42pm
I believe the “auction fever” phenomenon has been studied quite extensively in different environments, and there appears to be a fairly wide body of evidence that points to at least some emotional component. For example, a study that Ariely and Simonson did on electronic auctions back in 2003, showed that participants on average paid 15% more for items in an auction, and that 98% of them paid more than they would have if they had bought the same item at the fixed price. Another study found that in 42% of cases, the winner of the auction paid more than they would have using the fixed price option on the same site, which helps address the increased search cost theory.
I’m not going to link to it directly, because that seems to get my posts put into a moderation black hole, but if you search for “Understanding auction fever: a framework for emotional bidding” you’ll get a pretty good summary of the literature on auction fever, both from an economic and psychological point of view.
Jon Murphy
Jun 17 2019 at 1:08pm
No one denies “auction fever” isn’t a thing. The question on the table is whether it is the explanation, or even the primary explanation, for the behavior observed by Ariely’s correspondent.
David Henderson
Jun 17 2019 at 2:21pm
Thanks, Dylan. I would refer you, though, to Jon Murphy’s comment in response to you. I admit the possibility. It’s just that it’s shocking that he goes to his psychological explanation without even mentioning a straightforward one.
Dylan
Jun 17 2019 at 3:59pm
I’ll just note that in the quick read I’ve done of some of his academic papers on the subject, there is an explicit acknowledgement of the rational, economic perspective, and it seemed that they have designed their experiments in a way to try to distinguish between this and a more emotional response. I’ve not gone into the details of those methods at all, so I can’t vouch for how well I think they’ve actually controlled for this, but it is something he’s definitely studied. It makes sense to me that he wouldn’t want to go into that level of detail in a short piece on the WSJ, and instead just focus on the explanation that has come out of his and others’ experimental research.
Jon Murphy
Jun 17 2019 at 4:03pm
I don’t think his goal was to avoid “that level of detail.” If it were, he’d have gone for the simple explanation (the economic one) rather than the more complicated behavioral one. I think Henderson’s hypothesis is likely correct: he’s a victim of motivated reasoning. Again, this does not imply he is wrong. It just means he is likely missing a more simple, and thus likely more correct, explanation.
Dylan
Jun 18 2019 at 6:45am
What I’m saying is, they’ve actually done experiments that show that this simpler explanation does not appear to be the correct one, or at least does not the primary contributor to the behavior we actually see. They’ve got a lot of evidence, based upon information like people’s self reported motivation during an auction, to experimental data, to the kind of data I mentioned above about how auction participants tend to overpay. Some of that evidence comes from live auction settings like the one used in the question. They also look into William’s point below about getting more information from the behavior of the bidders and conclude that auction participants tend to overweight information they get from within the context of the auction, and underweight information from outside of it.
You may disagree with all of that, and think that the studies weren’t done well, or that the conclusions don’t follow from the data, or whatever. But, it does seem weird to me to criticize him for not talking about the simple explanation, when his research doesn’t support that hypothesis.
Charlie
Jun 18 2019 at 7:25am
It seems probable to me that both explanations play a part – the question under dispute seems to be which plays a stronger part.
Dylan says “What I’m saying is, they’ve actually done experiments that show that this simpler explanation does not appear to be the correct one, or at least does not the primary contributor to the behavior we actually see. […] But, it does seem weird to me to criticize him for not talking about the simple explanation, when his research doesn’t support that hypothesis.”
This is not the case as far as I can see. By your account :
“I’ll just note that in the quick read I’ve done of some of his academic papers on the subject, there is an explicit acknowledgement of the rational, economic perspective, and it seemed that they have designed their experiments in a way to try to distinguish between this and a more emotional response.”
This sounds like he controls for that hypothesis and finds some leftover effect which is due to psychological effects. This doesn’t mean his study claims the psychological effect is the stronger of the two.
The “motivated reasoning” John accuses him of is choosing, as a psychologist, to only emphasize the psychological explanation rather than the simpler and in John’s view (and mine, for what it’s worth) more powerful economic reason.
Jon Murphy
Jun 18 2019 at 12:24pm
Again, no one is arguing that “auction madness” is not a thing. It may, and probably does, play a role here. The question is whether it is the best explanation of the observation.
For any observation, there will be many explanations, all of which may have grains of truth in them. But what is the best of these explanations? Which are we going to rely on? Which are we going to use as a starting point when expressing ideas to the public?
Occam’s razor comes into play here. And, indeed, we see the explanations that require very few assumptions have a lot of power to them. Supply and Demand goes a very long way. It is not perfect, but it can tell us a lot. If I am talking to my grandmother and she askes me why some observation happens, I’m going to rely on straightforward supply and demand. I won’t burden her with ten trillion asterisks and subtleties, especially ones that require us to assume the mind of another person.
I realize that Ariely has a bais given his own work; that’s part of the point (see the last paragraph of the post). I have my own biases when discussing things related to my own work. We all do. That is an explanation for why Ariely misses the straightforward explanation of the observation.
Dylan
Jun 18 2019 at 1:31pm
We might be talking past one another here, and I’ve not done a thorough enough reading of Ariely’s work to know if I’ve interpreted his findings correctly, so take this as a hypothetical.
Let’s say I’ve done a bunch of research on this phenomenon. I’ve interviewed people before and after an auction, and before they’ve told me the top price they think an item is worth, and afterwards when they paid way more than that, they tell me the things they were thinking about had nothing to do with value, but instead about wanting to “win” the auction, and not let someone else get “their” item. Then I’ve run experiments in auctions to try and specifically control for the “endowment” effect (the idea that the act of bidding itself gives one a sense of ownership of the item, and therefor a higher value), and found that indeed, subject participants put in conditions where they are bidding over a longer period value the item more than participants who come in at later stages, even though they have the exact same information available to them. Then I look at real world examples over thousands and thousands of auctions, and find that the rational economic view doesn’t seem to explain what we actually see in these auctions. And given all this evidence, in my view auction fever does a much better job explaining what we see, then does the rational economic actor theory. Indeed, we’ve interviewed many, many people before auctions, and know that none of them initially values the item at above $20 (or at least they have told us that they don’t), yet the final bid nearly always ends up above that amount.
To me, this looks like the bulk of the evidence points to the “simpler*” explanation being wrong, or at least woefully incomplete. If you have a lot of space, I can see first taking some time to “debunk” the simple explanation, but if space is limited, just get right to the answer.
Dylan
Jun 18 2019 at 1:48pm
Simpler (or at least shorter) version. I think that Ariely’s work shows (or at least he thinks it shows) that the simpler explanation as given by Dr. Henderson is incorrect and does not explain the behavior of the auction participants. To reach this conclusion he doesn’t have to make any heroic assumptions beyond that the auction, and it’s participants, is not meaningfully different from the many that he has studied in the past.
Dylan
Jun 18 2019 at 5:17pm
@Charlie
Here’s the abstract from one paper (not Ariely). Apologies for not linking, but a Google Scholar search for the title will lead to the paper. Emphasis added to the part relevant to this discussion.
Towards a competitive arousal model of decision-making: A study of auction fever in live and Internet auctions
“In 1999, Chicago sponsored a public art exhibit of over 300 life-sized fiberglass cows that culminated in 140 Internet and live, in-person auctions. Collectively, the cows sold for almost seven times their initial estimates. These unexpectedly high final prices provided the impetus for a model of decision-making, “competitive arousal,” which focuses on how diverse factors such as rivalry, social facilitation, time pressure, and/or the uniqueness of being first can fuel arousal, which then impairs decision-making. In Study 1, live and Internet bidding and survey data from 21 auctions throughout North America tested the model’s predictions, as well as hypotheses derived from rational choice and escalation of commitment models. Analyses provided considerable support for the competitive arousal and escalation models, and no support for rational choice predictions. Study 2 was a laboratory experiment that investigated the similarities and differences between escalation and competitive arousal, finding again that both can result in overbidding. The discussion focuses on the implications of these findings and on the broader issue of competitive arousal and escalation and their impact on decision-making.”
Dylan
Jun 19 2019 at 8:38am
The research into this area is pretty interesting to me, so I’ve continued reading the studies in more detail. The study I quote above, (which is not an Ariely study, but is one that he’s cited in his research more than once, so I assume he’s familiar with it) explicitly looks at a scenario very similar to what David describes. The rational choice theory says that the presence of other bidders gives feedback about the value that others place on the object, and can provide a rational basis for exceeding the maximum limit an individual had set for themselves before the auction began. The researchers make a hypothesis that this theory is strengthened the more bidders are left near the end of the auction, i.e. if lots of people are valuing the object more than your maximum bid, this gives stronger evidence that the item is actually worth more than you thought. By contrast, the competitive arousal theory predicted that bidders would exceed their maximum limits when there were fewer bidders left at the end, which is what they saw in their studies.
Again, you may disagree with the conclusions that are drawn from these studies. I just think it is wrong to say that Ariely hasn’t considered the “simpler” rational choice explanation.
*In a post above I put an asterisk next to simpler, but forgot to include the note. I think the rational choice theory may seem like the simpler idea to economists, but I think for the average person, the “I got a little carried away by competition, and paid more than I should” seems like the more straightforward explanation.
Charlie
Jun 20 2019 at 11:06am
@Dylan. Yes, I think we’re talking past each other here. I think the particular case of “why does nobody bid $20?” does not need them to update their initial valuation at all, and is quite distinct from the rational choice theory described.
“The rational choice theory says that the presence of other bidders gives feedback about the value that others place on the object, and can provide a rational basis for exceeding the maximum limit an individual had set for themselves before the auction began. ”
This theory is not necessary for David’s argument, I think.. The only economic reasoning required to not bid $20 is they might get it for $15 if they wait.
It’s very possible that they all only valued it at $15 at the beginning, and fell victim to auction madness, but it’s equally possible they all valued it at $25, and thought they might get a better bargain by waiting. The rational choice theory vs auction madness debate is wholly unnecessary to explain not bidding $20 in the first place.
Dylan
Jun 20 2019 at 1:07pm
@Charlie
Thank you for the clear explanation, I’ve got a much better understanding now of the point that I think David was trying to make, and I agree we can’t really know the specifics of what is going on in this auction and why they didn’t bid at first. The question does state that the item “ended up selling for much more” than $20. Just to put some numbers here, let’s assume that the ultimate selling price was $100. To me, it seems if you value an item at $100, and you have the option to bid on it at $20, and no one else is bidding on it at that price, it would be smart to bid then and try to lock up the item. If you wait for the asking price to be brought down, you risk that more bidders will enter and you end up paying more than $20, which seems to be what happened here. The literature does appear to be consistent that lower starting bids generally lead to higher ending prices in auctions, so waiting for the opening bid to be lowered seems like a high risk strategy for not much gain, if you do value the item at a much higher level than the initial opening bid.
But still, I now see that while the situations are fairly similar, we don’t have enough information in this case to rule out either theory of motivation.
John Alcorn
Jun 17 2019 at 2:08pm
Sharp post!
A general point: When we try to explain a puzzling instance of behavior, we should (a) consider all plausible motivations and mechanisms and (b) check each possibility against the evidence. If the evidence isn’t sufficient to establish which motivation (or mechanism) was in play, then we should say so.
In the case described by Questioner Liron, game theory (David Henderson) and psychology (Dan Ariely) provide plausible hypotheses. However, we lack sufficient evidence to establish which hypothesis is correct in this particular case.
Moreover, there might be still other plausible motivations. For example, in particular settings there might be social norms about initial bids. We should always try and think of alternative explanations (conjectures), and then seek discriminating evidence.
Kevin Erdmann
Jun 17 2019 at 4:00pm
This is the danger with behavioral finance. And, worse, if the behavioral model is wrong, the modeler will tend to be emboldened by it. If my model says the price of something should be $10, and it’s $15, then the behavioral explanation is that some cognitive or behavioral bias caused a $5 anomaly. What if the price is $20, or $30, or $40. The behavioral observer will be tempted to conclude that the anomaly is even more strong and distorting than they had thought it was.
Working from other types of models, the more wrong your prediction is, the more you doubt your model. But, in behavioral econ, the more wrong it is, the more confident you feel.
To me this is the main problem with the responses to the housing bubble and the financial crisis. The credit-fueled bubble hypothesis was largely mistaken, but the more home prices continued to rise, the more emboldened so many observers felt in condemning the irrational and short-sighted agents that populated their models.
C Haller
Jun 17 2019 at 4:35pm
It’s comical the extent to which some of you will go to deny the fundamental irrationality of human beings.
I mean, just look around you!
The number of circumstances which truly reward what Daniel Kahneman calls “system 2 thinking” are rounding error in the lives of virtually all human beings in all places and times across recorded history.
There have been thousands of years of evolutionary pressure culminating to favor the use of heuristics in human decision making. Deal with it.
William Connolley
Jun 17 2019 at 6:11pm
Is there a third explanation…? If you’re bidding for a thing, perhaps you don’t know what it is worth. If no-one else bids $20, and you’re uncertain, you conclude that it isn’t worth that, and don’t bid. Perhaps you’d look stupid if you were the only bidder. But once other people start bidding, you have some extra knowledge – they value it too. Perhaps more true of things that are more aesthetic than functional.
David Henderson
Jun 17 2019 at 7:32pm
Interesting point.
MarkW
Jun 18 2019 at 8:31am
Yes, and you have to consider the possibility that the winning bidder is buying the item with the plan of reselling it at some point. In that case, the willingness-to-pay of other bidders is particularly useful information.
Justin Rietz
Jun 18 2019 at 3:18am
Piggy-backing off of William’s comment: an asset’s value is in part determined by its residual value. Therefore:
As the residual value increases – as indicated by other bidders revealing their willingness to pay, and therefore indicating the amount for which I presumably could turn around and sell the asset – my valuation of the asset rationally increases.
Knowing the residual value with increased certainty (e.g. having multiple bidders) rationally increases my valuation of the asset as future risk is reduced.
john hare
Jun 18 2019 at 4:56pm
I seem to be missing the point of the post and most commenters. On vehicles and equipment, we consistently pay well under retail value according to KBB. A bit above trade in value, but not much It is assumed that if the vehicle is at auction, there are problems that will have to be addressed before putting it on the road. We discount what we will bid based on perceived risk. Only during tax refund season do we see people paying above retail.
All this talk about paying more at auctions than the item is worth is very murky to me. Is this the personal experience of people commenting here?
David Henderson
Jun 18 2019 at 7:33pm
You wrote:
I seem to be missing the point of the post and most commenters.
I’m surprised that you say you’re missing the point of the post. I don’t know how to make it clearer other than to say that you might want to read the first comment, the one by Jon Murphy, that makes my point better than I made it. Take a look at it again and see if you still are missing the point.
john hare
Jun 18 2019 at 8:22pm
Perhaps I said it wrong. You and Jon were very clear in your writing. It’s just that I see a different outcome in the auctions I attend. My friends and I get bargains at auction compared to street value. I don’t see the overbidding described here. Even the tax return wealthy are paying less than they would off of a car lot.
So I’m understanding what is being written, but it’s contradictory to what I experience. Perhaps it’s the type of items being sold.
James
Jun 18 2019 at 9:23pm
“All this talk about paying more at auctions than the item is worth is very murky to me. Is this the personal experience of people commenting here?”
There is an economic theory (winner’s curse) that the winner of an auction will generally pay more than the item at auction is worth. For example, say I am auctioning a bag of change that is actually worth $100. I know exactly what’s in the bag but I keep this information to myself. Every bidder will have an estimate in mind of the actual value. If these bidders are unbiased as a group, half of these estimates will be too high and half will be too low but the average estimate will be $100. So who will win the auction? It will be the bidder who overestimates the most.
When people do lab experiments with randomly selected people it works just like the theory says. That is not in dispute. So why can’t we just get rich auctioning bags of change?
In real life, people who are bad at stuff generally stop doing what they are bad at, either because they realize they are bad or because they conclude that the process is rigged or because they run out of money. The people who are good at bidding in auctions will have positive experiences (like yours) and continue to go to auctions. The people who are bad at auctions will eventually give up on auctions. That’s why at most auctions, the majority of bidders will be people with a history of finding bargains at auctions.
John Alcorn
Jun 19 2019 at 11:51am
Hal Varian discusses behavior at auctions in his conversation with Tyler Cowen (at time cue 00:09:45-00:11:30).
Dr. Varian indicates that auction fever (Ariely’s mechanism) and the winner’s curse (noted by James in his comment above) have some bite, especially among relatively inexperienced bidders:
john hare
Jun 20 2019 at 6:16pm
@Dylan 1:07 pm.
My experience is at variance with these studies so this may not work for others.
The final bid is roughly in proportion to the starting bid. Seldom does it exceed triple the starting bid. I risk nothing by waiting for the lowest possible start price. If I bid $20.00 then I am sending the signal that the item is worth at least that. It could easily go to $50.00-$60.00. If I wait for the $10.00 start, it is unlikely to exceed $30.00, and I may get it for $10.00.
Dylan
Jun 20 2019 at 6:36pm
I’ve got a feeling that this can vary a lot depending on the type of auction and setting. We used to have an auction house down the street, and the kinds of auction they did, business liquidation, overstock, some government, etc. I’m guessing that is more similar to the kind of auctions you attend, and I’m not sure those are similar enough to the ones that have been studied to make comparisons. Perhaps the search cost, inconvenience, and not really knowing what your getting make those auctions significantly different from both eBay and in person collectibles and art auctions, along with experimental auctions done in an academic setting?
My only personal experience with an auction was on the set of a TV show I really liked that had been cancelled, and I ended up walking away with a couple things I didn’t really want, just because I wanted to have something from the show. So I’m not the best judge.
Comments are closed.