But here we are in 2012 and all that has changed. We have seen the unthinkable occur, again and again: a major investment bank going bankrupt, housing prices plunging across the board and staying down, a global near financial meltdown, people actually paying serious attention to Nassim Taleb – and now we face the  imminent possibility of default on sovereign debt by major European nations, and perhaps even the collapse of the Euro.

The topic now seems, if anything, too timely. Every second article in the financial press seems to be about when the next break in the global economy will come, how far it will spread and how rapidly. Rest assured, this is not yet another Chicken Little post. The economic sky may in fact fall, but that’s for others to predict.

What I’m interested in here is: What actually happens with individual investors that results in panic, contagion and mass market movements, and how can we spot it before it comes crashing down around us?

Let me be clear: by “contagion” I am not referring to the truly terrifying structural linkages by which, say, default by Greece would inevitably take down numerous European banks and probably a few other countries. These are “hard-wired” into the system and only require occurrence of the initial unthinkable event to set them off. What I am interested in exploring is the “psychological contagion” that looks like a mass panic, where for example the failure of one well-regarded bank leads to runs on other, sound, banks, or a drop in certain stocks leads, not to the expected offset by opportunistic buyers, but to an “unreasoning” stampede out of stocks in general. How can we make sense out of these seemingly senseless mass market movements? Let’s begin by marking out the territory to be explored.

Mass-market  panics:

1. Start with a sharp break from prior market behavior. It looks as if, across the board, something important changes, which results in markedly different behavior.
2. Spread rapidly, like flu epidemics or panicked stampedes in burning buildings.
3. Are essentially impervious to reason, reassurance and cold, hard facts.

Some obvious questions: What is the important change that results in the sharp break in behavior? How, exactly, does the contagion spread? And why, oh why, do people ignore informed pleas for reason as they figuratively stampede to the nearest exit?

The usual candidate for “what changes” is: market sentiment. But that’s just a label, not an explanation – one day “the market” sees cautious opportunity, the next day it sees massive risk. Why? And why does it happen so abruptly?

We have even less understanding of how the “contagion” spreads into “panic”. Note the quotation marks: they are intended to remind us that “contagion” and “panic” in this case are metaphors, borrowed from realms in which they are actual facts. Diseases are in fact contagious and we know both what causes them and the biological mechanisms by which they spread. We know nothing at all comparable about market “contagion”. Likewise, panics actually happen in crowds when someone perceives danger, takes strong action to avoid it and others react to the danger-avoiding behavior. Physical proximity is an essential part of the spread of panic; there is no physical proximity in market panics except, for example, among traders jammed together on a trading floor. The puzzling thing about market “panics” is precisely that there is no physical proximity and therefore no plausible explanation for how the “panic” spreads.

Finally, the “explanations” for why informed pleas for reason are ignored simply say that this is emotional, therefore irrational, behavior. This is the standard dismissive fallback of economists (even behavioral economists) and as we have seen in prior posts, it completely misses the point. Behavior makes sense; saying that it doesn’t just means you don’t know where to look to understand it. We can do better with all three questions because, as Descriptive Psychologists, we know where and how to look.

We start by restating and then expanding one of Descriptive Psychology’s core formulations about people and behavior:

A person acts on the state-of-affairs he currently finds himself to be in. (“Finds himself” does not imply passivity; the current state-of-affairs routinely is the result of the person’s prior intentional actions.) Certain things are actual parts of the current state-of-affairs; they are real to the person when she acts. Other things are potential parts of the state-of-affairs; they are possible to her when she acts. People value both the real and the possible, and they value them differently. Specifically, a person will value a real something over the “same” possible something.

To this we add the third and final piece of this formulation:

Possible parts of the current state-of-affairs divide into two categories: things the person takes to be actually possible, and things she takes to be  merely theoretically possible. A person will take actually possible things into account in acting; he will not take into account things that are merely theoretically possible.

So what is the important change that results in the sharp break in behavior? Just this:

Something changes from merely theoretically possible to actually possible for this person.

And that something – whatever it is – is now taken into account in the person’s actions, whereas before it was not.

Note a few things about this formulation:

• The change from theoretically possible to actually possible is a categorical change – not a matter of degree, but of kind.
• This change is a state change. Just as state changes in quantum physics happen instantaneously and without an intervening process, so too do changes from theoretically possible to actually possible.
• There is no room at all here for talk about “probability”. Either a thing is actually possible to a person, or it is not.

The near-meltdown of the financial system in 2008 occurred when, for many investors, bankruptcy of a major investment bank changed from merely theoretically possible to actually possible. What made it actually possible was a real occurrence – it actually happened to Lehman – and therefore could possibly happen again to other banks. From that point on, the possibility of major bank failure was actually considered by the investors – and for a large majority of investors the actual possibility of bank failure is too great a risk to countenance. Runs on (and collapse of) other banks would have been inevitable had not the U.S. government taken extraordinary (indeed, unprecedented) measures to assure investors that another bank failure would not be allowed to happen – and that they had the will and financial muscle to back up their assurances.

Notice that there is no place in this account for talk about emotion or irrational behavior. A person may very well feel a cold lump of fear on realizing that they have been ignoring a large and important risk, and they may experience shock when they suddenly realize things are different than they thought – but those are just feelings that accompany a rational assessment of the situation. A person acts – sensibly – on what they take to be the case, not essentially on what feelings accompany their assessment.

How does this spread? Not primarily from person to person, but systematically.  Simply, it’s the assessment of what is actually possible that spreads through the market, as more and more individuals discover they have reason to reevaluate those “merely theoretically possible” things. The rich stew of information, stories, rumors, and prognostications that form our daily media fare is available to everyone, and while insiders usually know about risks before the rest of us, eventually something happens that tips the scales: The impossible or theoretically possible becomes an actual possibility, and then behavior changes rapidly, en masse – as if a panic were occurring or an epidemic spreading. But of course it is actually a large number of people reaching a state-changing conclusion within days or even hours of each other.

This formulation leads to some important further questions:

• Why are these market “panics” so resistant to facts and reason?
• What does it take to make the theoretically possible actually possible?
• Once something becomes actually possible for a person, does it stay that way?
• Can we spot when a state change in the market is about to occur, and can we do anything about it?

Answering these questions will require a bit more conceptualization from Descriptive Psychology. We will take these up in the next post in this series.

One parting note: Consider a recent statement from the New York Times:

“Markets, which had always assumed that major Western governments would honor their obligations, struggled to learn to adjust to a new world where that was not so certain.” The Year Governments Lost Their Credibility, NY Times 12/30/11

Has “major Western government” default become actually possible in the market’s eyes– or are we still just flirting with a theoretical possibility? Perhaps government bonds currently are, as British slang once put it, “safe as houses”. That’s an ironically unsettling thought.

What are we to do about the fact that people value what they have over the same thing belonging to someone else? About the fact that buyers of stock are overly reluctant to sell it at a loss, and that champions of projects continue to fund them when clear-eyed analysis says that pulling the plug is the better course? We begin by recognizing, as previous posts demonstrate, that such behavior is not “irrational”; in fact, in light of the person’s circumstances, the behavior makes sense. Once we see the sense it makes, we’re in a position to do something sensible about it.

Let’s review our analysis of John, the stock buyer from “What’s Really Going on With Sunk Cost?” John buys a stock for $100.00 per share. Eight months later, the stock is down to $80.00 per share.  Does he give the stock time to come back, or does he sell it?

Prior to selling his stock for $80.00 per share, two things are true for John:

1. He has stock instead of the equivalent amount of cash.
2. He has a potential loss of $20.00 per share.

After John sells, two things are true for him:

1. He now has cash instead of the equivalent amount of stock.
2. He has actually lost $20.00 per share.

In other words, selling leads to an actual economic loss, where not selling leads to a possible loss – and a person will value a possible loss over the same actual loss. We would predict that John would not sell more often than he will sell, which of course is exactly what the research data show. Further, John’s state-of-affairs also includes matters that are not directly wealth-related – for instance, matters of standing. Selling at a loss turns a potential hit to his image as a successful trader to an actual hit. So John has two potential losses to weigh against the same two actual losses – a strong motivation to hold rather than sell.

Now let’s suppose John works for you, as part of a trading enterprise. You want the buy/hold decision to be made based solely on the merits of the transaction, ignoring sunk cost. How can you make this more likely to happen?

Well, you could train John on issues of sunk cost, and make it an operating principle to ignore sunk cost in these decisions, and sometimes this works. In fact, you’ve probably already done that. And you can insist that professional traders take a clear-eyed view, knowing when to hold and knowing when to fold, and that cutting John’s losses will not reflect badly on him; again, this may assure John that his standing will not suffer in your eyes if he sells this stock. (But hold on a minute: Is that in fact true in your shop? Don’t you have metrics that track how individual traders do, and won’t this sale-at-loss in fact show up negatively on those metrics? Do you hear senior people boasting about cutting their losses, or do you just hear about their big winners? If so, why would John believe the official line when day-to-day reality shows otherwise?)

John is a person, not an algorithm, so despite the training and admonitions, he still has reason to avoid the hit to his standing as a successful trader, both in your eyes and his own, and he is likely to act on those reasons.

What you can do is: Take the hold/sell decision out of John’s hands and give it to someone else. Mary does not have John’s actual/potential losses associated with this stock, so she is more likely to take that clear-eyed view you need. You can correct for the “endowment effect” bias with a simple operating principle: The person who makes a buy decision does not make the hold/sell decision.

Putting this principle into practice requires a little care to avoid creating perverse incentives (e.g. Mary and John are competitors for bonus and advancement, so Mary has reason to tag John with a loss if she can, and sell off prematurely what might have been a good winner for him). But competent care is all it takes to in fact structure the trading environment to avoid “endowment effect” bias.

A note for investment advisors: If you switch from John-working-for-you to you-advising-John-on-trading-for-his-own-account, John is even more likely to act to protect his standing in his own eyes as a successful trader. Perhaps you would serve your clients best by contracting with them to keep a close eye on any investment they have made and actively advocate for selling when the time is right. Assuming they are not complete beginners, they probably need you more for the sell decision than for the buy.

And finally, what if you are John or Meg – a person who makes your own investments decisions for your own account? Instead of employing a professional money manager to invest for you, you subscribe to some research or data services and decide for yourself how much of what to buy and when. How can you avoid getting caught in valuing what you own higher than you should?

Try the classic “stop-loss” play: When you decide to buy, before you place the order, decide at what price you will sell to get out of a losing position. When you buy, at the same time instruct your broker to sell if the price drops to your limit (if you are trading from an on-line account, set the limit yourself) and then do not revisit your decision. Once you own a stock, you have reasons to keep it that have nothing to do with its actual current value – and although you may think you can be objective, the chances of that are actually quite slim.

Here’s to taking the cash and letting the credit go in 2012!

It is a standard accounting term, referring to the investment already made in an asset or project. When deciding how to value the asset currently, or whether to invest further in the project, we are firmly admonished to ignore “sunk cost”, and to make our evaluations based on its current market value (or expected future value). This is sound advice, rational to the core, and is perfectly in accord with classic economic theory.

So what’s the rub? To the consternation of accountants, economists and financial advisors everywhere, when we look at actual investment decisions, people routinely do NOT ignore sunk costs. What they have already invested is typically a factor – sometimes a determining factor – in what they decide to invest now. Talk about “irrational”! And since behavioral economists do talk about “irrational” they account for this and many other common deviations from rational practice by invoking the “Endowment Effect.”

In the first post of this series, “Take the Cash and Let the Credit Go”, we saw how “loss aversion” is an ad-hoc account of research data that merely labels the phenomenon but in no way explains it, and gave an alternative formulation that actually predicts the research findings. In this post we will subject the “Endowment Effect” to the same treatment, with an interesting twist: exactly the same conceptual structure that predicts loss aversion also predicts the Endowment Effect! This is our first solid clue that we are on to something really different and powerful here. (And there’s a good deal more where these come from, as we shall see as this series unfolds.)

A large body of research has established an interesting fact: if people own  something, they value it more than they value that same thing belonging to someone else. And if they have gone to some trouble to acquire the thing, such as standing in line for a long time or making it themselves, the difference in valuation can sometimes be dramatically large.

Consider one of Dan Ariely’s studies, in which he found that Duke students who won tickets to basketball games via a lottery asked on average $2400.00 for their ticket, while other Duke students who did not win that lottery offered an average of $170.00 for a ticket – in fact, no ticketholder was willing to sell for what someone was willing to pay! In his  2008 book Predictably Irrational Ariely accounts for the large difference in terms of emotional attachment and “imagined glory of the game.”

Behavioral economists call this the Endowment Effect, saying essentially that owners endow the things they own with an irrational extra value, emotional in origin, that distorts their valuation of the thing. Once again, notice that we have an ad-hoc label, applied to research findings, that explains nothing. We can do better than that, and we shall.

Consider the classic “sunk cost” conundrum for investors:

• John buys a stock for $100.00 per share. Eight months later, the stock is down to $80.00 per share.  Does he give the stock time to come back, or does he sell it?

Or the “sunk cost” conundrum for executives:

•  Meg launched a new product 12 months ago, in which her company has invested heavily. Sales have been substantially lower than forecasted and the market seems lukewarm to even the latest model. Should she stay the course and give the product a chance to take off, or should she cut her losses and close down the project?

In both cases, the “rational” approach would be to ignore the sunk cost, and make the decision on the basis of the best current information and estimates. We should ask: “Would I buy this stock now if I didn’t already own it?” “Would I fund this project if someone brought it to me now for the first time, with all the product and market data to date?” and act accordingly. This is not exactly breaking news; this is investing 101. But the simple fact is: John will hold onto that stock he owns far more often than he would buy it now for the first time; Meg will extend the project far more often than she would fund it fresh; and this applies not just to amateurs – professional investors “err” in the same direction.

At this point we can either sound the common lament about the irrationality of people who get emotionally attached to their investments, or we can acknowledge that more is going on here than we have considered. Let’s try that less-traveled road and see where it leads us.

Recall from the previous post:

•  A person acts on the state-of-affairs she currently finds herself to be in. (“Finds herself” does not imply passivity; the current state-of-affairs routinely is the result on the person’s prior intentional actions.) Certain things are actual parts of the current state-of-affairs; they are real to the person when he acts. Other things are potential parts of the state-of-affairs; they are possible to him when he acts. People value both the real and the possible, and they value them differently. Specifically, a person will value a real something over the “same” possible something.

As a corollary to this:  A person will value a possible loss of something over the same real loss.

Prior to selling his stock for $80.00 per share, two things are true for John:

1. He has stock instead of the equivalent amount of cash.
2. He has a potential loss of $20.00 per share.

After John sells, two things are true for him:

1. He now has cash instead of the equivalent amount of stock.
2. He has actually lost $20.00 per share.

In other words, selling leads to an actual loss, where not selling leads to a possible loss. We would predict that John would not sell more often than he will sell, which of course is exactly what the research data show. And notice that we don’t need to invoke some emotionally-based “endowment effect” to understand what’s going on; John’s behavior is fully explained by the observable facts of his situation.

Note also that we have taken John’s to be a “pure” economic transaction – nothing is considered other than the change in “states of wealth”, as Kahneman and Tversky put it. To expand a bit: the “state of affairs” John finds himself to be in both before and after his decision, is taken to consist of nothing but matters of wealth. This does, of course, happen in real life – but it is fact relatively rare. Far more often John’s state-of-affairs also includes matters that are not directly wealth-related – for instance, matters of standing. And when we look at that picture, we find that “sunk cost” is even less relevant to his decision.

Consider Meg’s dilemma. When Meg launched the new product, she put two things on the line: the assets of her company, and her own standing as the A-list executive leading the project. One year later she is an A-list executive whose most recent project is a possible failure. If she continues the project, her standing remains as it is. If she cancels the project, her standing becomes an A-list executive whose most recent project actually failed.

Now Meg has two possible losses: the loss of her company’s investment, and the loss to her standing, to weigh against the same two actual losses. The case for deciding to continue is now stronger than when we considered only wealth matters – as a Descriptive Psychology Maxim points out, “When a person has two reasons to engage in a particular behavior, he has stronger reason to act than when he has only one of those reasons.” We now would predict that Meg will continue the project significantly more often than she will cancel. And again, note that there is nothing emotional or irrational here; Meg’s behavior makes sense in light of her circumstances.

In fact, let’s make that a general statement:

People are not “rational” utility-maximizers as classic economics insist, but neither are they the “irrational” victims of cognitive biases, as behavioral economists say. People make sense – as people – and their behavior makes sense,  as behavior.

That’s what Descriptive Psychologists say.

For more, see the academic paper Joe Jeffrey and I wrote, which will be published in an upcoming issue of the Journal of Behavioral Finance http://www.descriptivepsychologyinstitute.org/IrrationalityIllusion.pdf

Next: An early present for the New Year. Perhaps surprisingly, this way of understanding leads directly to some practical guidance for how corporations and investors can do a better job in light of the actual “sunk cost” dilemma. The next post will give some of that guidance.

Solving the oil shale problem seems to require better and stronger technology; solving the behavioral science research problem requires better and stronger conceptualization. Fortunately, that conceptualization is available and well-developed, in the powerful conceptual net called Descriptive Psychology. What is called for now is a thorough reconsideration of behavioral science research, to create a fresh understanding of what has been discovered that unlocks the value in the findings.

Let me be clear: since at least the 1970’s an enormous body of sound empirical research has been accomplished in behavioral science, including, among many other domains, behavioral economics, social psychology, cognitive psychology and neuroscience. I take these findings as given. It is not my intention to question the findings; rather I intend to question the explanations researchers have given of their findings, and to offer other, more plausible and considerably more powerful explanations in their place.

As Omar Khayyam said in a different context, I intend to “Take the cash, and let the credit go.”

Here’s a specific example: Daniel Kahneman is a renowned cognitive psychologist whose works stand in both professional and popular literature as empirically verified truth. He is cited, with his research partner Amos Tversky, as a founder of behavioral economics. Awarded the Nobel Prize in economics in 2002, his findings and formulations have permeated the professional literature and are widely reported in popular accounts of modern behavioral science. Questioning Kahneman’s findings is a bit like questioning Gandhi’s ethics – how would one dare, and who are you to question him, anyway?

Daunting, true, and made even more so by my personal admiration for Kahneman’s work. There is no doubt that he has accomplished a great deal of lasting value. What is in doubt is exactly what he has accomplished.

The empirical work itself is top-notch, but his interpretation of it – his explanations of what it shows – are not. From the beginning, Kahneman’s account of his research findings has been deeply and systematically flawed. Let’s look at an example from his earliest publications to clarify the problem, identify the root cause, and illustrate the difference an alternative explanation can make.

In 1974 Kahneman and Tversky published an epochal paper reporting on a fascinating set of findings. In essence, here’s what they found:

Condition 1: Give one group of people $10.00, and then require them to perform a task to a certain level in order to keep the money. Condition 2: Promise another group of people that if they perform that same task to the same level, you will give them $10.00. Exactly the same task with the same outcome; classic economics unequivocally predicts that people will try equally hard in both circumstances. The actual results? People routinely, consistently, predictably try harder in Condition 1 than in Condition 2.

These are fascinating, important findings. For one thing, they flatly contradict the prediction of classic economics and require rethinking the entire model – which Kahneman, et. al. did, to good effect, calling their new model “behavioral economics.”

But perhaps more interesting: unlike much of behavioral science, this doesn’t seem like merely reconfirming common sense with research subjects. Most of us would not have predicted this outcome; we certainly would not have predicted the decades of studies, in many contexts, that again and again confirm the findings. There’s something real here, not intuitively obvious, and predictably consistent; it calls out for an explanation.

And of course Kahneman and Tversky offered an explanation, which has become as close to established truth as you can find anywhere:

People suffer from “loss aversion.”

Why do people work harder to keep what they have than they will work to get the same thing? Why, people have “loss aversion”, that’s why. And that’s been the story ever since 1974; people have “loss aversion” and can be predictably dealt with on that basis. This idea has become accepted as simple fact by behavioral scientists and behavioral economists.

What’s wrong with that? Please recall, I have no issue with the findings; again and again this interesting fact about people has been demonstrated. The issue with is the “loss aversion” explanation of the findings. As the founder of Descriptive Psychology, Peter G. Ossorio, once said in a different context: “If you were buying explanations on the open market, you wouldn’t pay a nickel for that one.”

“Loss aversion” fails the critical tests for scientific explanation because it is ad hoc. It was created after the fact specifically to fit the research findings. Kahneman and Tversky did not say, as empirical science requires: “We have a view or theory of human behavior that leads us to predict ‘loss aversion’, and here is our experimental design that tests our hypothesis.” Instead, they looked at the empirical results, noticed a significant pattern and called it “loss aversion”. But by doing so they have explained nothing whatsoever; they have merely provided a label for their findings that sheds no light on what the findings might mean.

It’s as if we had examined a large number of corpses, noted a specific pattern, and concluded that people under certain predictable circumstances suffer from rigor mortis. Well, they do, but that doesn’t explain what happens; it just provides a label for it. To understand what’s going on with rigor mortis we have to understand a great deal about how human bodies function. Then we can use rigor mortis to understand other important things, such as time of death. “Loss aversion”, like all ad hoc explanations, cannot be used to understand anything further.

If this were the only instance of flawed explanation, it would not be worth getting worked up about. But as we look a little closer, we find that all of Kahneman’s explanations – indeed, virtually all explanations to be found in behavioral economics and much of the rest of behavioral science – suffer from this same flaw: They are ad hoc, merely labeling the findings, and offer no further explanatory power whatsoever. Further, as one direct result of this state of affairs, the findings of behavioral research have not cumulated into a complex, interconnected picture of people and behavior, as one might expect; instead, they have created a large number of “insights” that have little discernible relationship to each other. Again, it’s as if we know a great deal about rigor mortis, skin diseases and spinal alignment, but nothing in particular about the human body as a whole.

Now this is an unexpected state of affairs that calls out for an explanation. What accounts for this overwhelming preponderance of ad hoc “explanations”? The root cause is obvious: none of these researchers are working from a comprehensive view or theory of people and behavior that could lead to testable hypotheses. In fact, for decades it has been taken as given that no such view exists, so researchers have done what they could without it.

That comprehensive view exists in the conceptual net called Descriptive Psychology. It has been articulated, developed and proven through both practice and research over the past 50 years. Let’s see what sense we can make of Kahneman’s “loss aversion” findings when we look at them from the viewpoint of Descriptive Psychology:

A person acts on the state-of-affairs she currently finds herself to be in. (“Finds herself” does not imply passivity; the current state-of-affairs routinely is the result on the person’s prior intentional actions.) Certain things are actual parts of the current state-of-affairs; they are real to the person when he acts. Other things are potential parts of the state-of-affairs; they are possible to him when he acts. People value both the real and the possible, and they value them differently. Specifically, a person will value a real something over the “same” possible something.

In condition 1 of Kahneman’s experiment, when the person performs the task the $10.00 is already theirs; it is real to them. In condition 2, when the person performs the task the $10.00 is not yet theirs; it is possible, but not actual. Based on this, Descriptive Psychology would predict that people would work harder in condition 1 than in condition 2 – and of course, that is exactly what happens.

In other words, loss is not stronger than gain; real is more valuable than possible.

Notice that this is not an ad hoc explanation like “loss aversion” that just happens to fit the data; it is a prediction derived from a comprehensive conceptual framework regarding people and their behavior. This is how empirical science in meant to work. And in the next post in this series, on “sunk cost”, we will see how it also meets the second test: It connects to other findings to cumulate into a complex, interconnected picture of people and behavior.

In this series of posts (“Take the Cash and Let the Credit Go”) in 2012 I intend to examine a number of the more important ad hoc explanations of findings in behavioral science research, beginning with more of Kahneman. Actually, a good start on this undertaking has already been made in an academic paper Joe Jeffrey and I wrote, which will be published in an upcoming issue of the Journal of Behavioral Finance – here’s a link:
http://www.descriptivepsychologyinstitute.org/IrrationalityIllusion.pdf

Any and all comments and feedback are welcome. Thank you for reading.

Well, OK. You could put it that way. But in doing so you are entirely distorting the reality of what is going on (and of course that was the Captain’s intention.) Unfortunately economists seem unaware of just how badly their talk about “incentives” distorts the realities they are talking about.

Case in point: Dan Ariely, in his new best-seller The Upside of Irrationality, cites a famous 1908 rat study by Yerkes and Dodson. Rats learned a maze slowly when the floor gave them a mild electrical shock (the shock stopped when they solved the maze); they learned the maze quicker when the shock increased to moderate levels; but when the shock became very intense, the rats performed worse.

Economists consider electrical shock (or anything negative) as an incentive to act. From this, economists concluded that the Yerkes and Dodson experiments demonstrate that incentives are related to performance in an inverted U-shaped curve: Up to a point, increasing incentives increases performance. Past that point, increasing incentives decreases performance. Furthermore, economists consider that seeking a bonus (or anything positive) is also an incentive to act; having re-described the experimental results in terms of incentives, they assert that bonuses are also related to performance in an inverted U-shaped curve.

So Ariely set out to perform experiments on people that paralleled the rat study, this time using bonuses as incentives rather than electrical shocks. And – surprise! – he found the same inverted U-shaped curve.

I have no issue with the results of either the rat study or Ariely’s human versions – they are interesting and call for explanation. (I admit to a queasy feeling about the ethics of Ariely’s experiment, but that’s a separate matter.) But the “incentives” explanation economists give is a bit like “a failure to communicate” – you could say that, but in doing so you are badly missing the point.

Let’s skip the poor rats and go directly to Ariely’s bonus incentives experiments with human subjects (he describes his methodology in some detail in his book, pp. 22-29). Participants performed 6 tasks requiring concentration; some also required manual dexterity while others required mental agility. They were informed beforehand that they will receive rewards – “bonuses” – for performing each task at a good level, or at a very good level (specifically defined for each task) and are told what their reward will be for good or very good performance.

The participants were assigned to one of three “bonus” levels: low, medium and high. Ariely meant the experimental results to throw some light on executive bonuses for CEO’s in business; accordingly he set the bonus levels to roughly equal one day’s pay (low); two weeks pay (medium); or 5 months pay (high). And since he couldn’t possibly afford to pay any participant 5 months pay in the USA, he took his experiment to rural India where the average daily wage was about 17 rupees (37 US cents).

So here’s the picture: in a rural village in India a researcher asks a passer-by if he wants to earn some money by playing some games. The games are demonstrated, the rules laid out and the stakes are announced; then the villager does his very best and earns whatever bonus he earns.

And the results? Interestingly, it was not an inverse U curve. Participants in the low bonus condition performed at a certain level (and generally left happy to have made a few extra rupees); participants in the medium bonus condition performed about the same (and left ecstatic at their good fortune); and the participants in the high bonus category – the ones who could earn a bonus of 2400 rupees, five months wages – performed significantly worse than either category. They left – well, let’s let Ariely tell that story:

“Anoopam … grabbed the two knobs that controlled the tilt of the maze frame and stared at the steel ball at its ‘start’ position as if it were prey. ‘This is very, very important,’ he mumbled. ‘I must succeed.’ … Unable to control the fine movement of his hands, he failed time after time. … With two more games to go [having earned only one ‘good’ bonus of 200 rupees] Anoopam decided to take a short break. He went through a calming breathing exercise, exhaling a long “Om” with each breath. … Unfortunately, he failed both … As he left the community center, he comforted himself with the thought of the 200 rupees he had earned – a nice sum for a few games – but his frustration at not having gotten the larger sum was evident in his furrowed brow.” And that was by no means the worst case: “[One] poor fellow was so nervous that he shook the whole time and couldn’t concentrate.” He got nothing.

The rats in the maze had it easy by comparison.

Ariely’s conclusion? See, just as we thought: big bonuses do not in fact increase performance.

Really? What we have here is … a failure to incentivize? With respect, I suggest that conclusion is far off the mark (and in the course of reaching those conclusions Ariely’s researchers brutalized some poor people about as badly as the guards brutalized Luke.)

What we actually have here, I suggest, is the fallacy of the Imperial Researcher: It’s my experiment, so what’s going on here is what I say is going on, and I say what the results are. The experiment shows that high levels of incentive decreases performance, and that’s that.

There are other (and better) explanations of what was going on here that do justice to the facts as we know them, and lead to very different conclusions. Let’s go back and look at the experiment from a somewhat different angle.

Ariely views his experiment as a classic form: ask people to engage in a specific behavior (playing some games); vary one independent variable (performance incentives); measure the dependent variable (their actual performance); plot their performance against their incentives.

Now notice something simple and obvious: the experiment depends on each person engaging in the same behavior. If one person is timed while solving a crossword puzzle while another is timed for a mile run, it makes no sense to compare their times and conclude the difference stems from their incentive. But that is exactly what Ariely has done.

Hold on, you might say – the people in India all engaged in the same behavior: playing some games (that is explicitly Ariely’s view.) But did they, really?

Consider: Rob points a gun and pulls the trigger. Bob also points a gun and pulls the trigger. Both engaged in the same behavior, right?

But now I tell you Rob was shooting at a target in a shooting range, while Bob was shooting at a burglar standing 10 feet away. Are you still inclined to say they engaged in the same behavior? Of course not. Rob and Bob engaged in the same overt performance (aiming and shooting a gun), but what they were actually doing by shooting the gun was very different.

Here’s the essential point: Ariely, along with virtually all behavioral researchers these days, explicitly takes the overt performance to be the behavior, even when it clearly is not.

In India the overt performance of playing some games was obviously not what the participants were doing. To understand what was really going on, it’s helpful to have some cultural reminders. People who live in rural Indian villages live at best a hand-to-mouth existence, and they virtually never have any opportunity to change their situation for the better. To have the chance to make a day’s wages by playing some games is great good fortune. To have the chance to earn two weeks wages in one day is amazing, unprecedented, and can, for a while, remove the daily pressure to earn today or go hungry.

But to have the chance to earn 5 months wages in a day: that is unimaginable; that is permanently life-changing. With that kind of money you can set yourself up in a small business and move out of daily labor for wages. You will never have an opportunity remotely like this in your lifetime, and you know it.

So what are the Indian villagers actually doing by engaging in the overt performance of playing those games? The low-incentive group is taking a shot at getting a day’s wages – something they do every day. If they succeed, great; if not, no big loss. But the high-incentive group is taking the one chance they will ever get to significantly change their life for the better. If they succeed, they will be in a different world; if not, they will have blown the only chance they will ever get, and they will probably regret it for the rest of their lives.

So – would you say the two groups are engaging in the same actual behavior? Of course not. Ariely’s experiment amounts to showing that when people engage in different behaviors, their performance may well be different. Add to that a well-known fact from sports: When a lot is riding on the outcome, it can be much harder to perform a task requiring concentration and physical or mental precision. (The putt you can make easily in a practice round is a lot less likely to fall when it’s to win the US Open.) That’s pretty much all these experiments “show”.

Here’s an irony: I suspect Ariely’s basic premise – that huge bonuses do not in fact improve CEO performance – is essentially correct, even though his experiment has no bearing on the question. But then it’s rather naïve to expect that huge bonuses would. Boards give huge CEO bonuses for the same reason lottery commissions happily give out huge prizes: for the effect they have on people’s willingness to play. The CEO may not work longer or harder because of the bonus, but you can be sure that thousands of other people in the company are inspired to work longer and harder to earn their shot becoming a CEO.

So are huge CEO bonuses “irrational”? Only in the narrowest possible view, which is not a view that includes the real world. Unfortunately, that view is typical of behavioral science as practiced today. Isn’t it time to change that?

The critique is of the work, not the man.

I admire the creativity of Ariely’s experiments and his passion to get at empirical truth. I have nothing against him personally – in fact, we have never met. In his books he comes across as an interesting and essentially nice guy; if we sat down for a cup of coffee I suspect I would enjoy his company and conversation.

Furthermore, I like that his research and critique consistently undermines the absurdity of classic economics and its “rational” homo economicus. For example:

“After all a truly rational person would generally not spend any money on anything or anyone that would not produce a tangible return on investment.” (Upside, p.247.)

And: “Simply put, it turned out to be extremely difficult for participants to think about calculation, statistical information and numbers, and to feel emotion at the same time.” (Upside, p.248.)

Good stuff, and there’s a lot more.

Unfortunately, there is also much not to like in Ariely’s work and that’s what the series is about. Ariely does interesting experiments, but his conclusions are consistently flawed and even misleading. This is a problem with behavioral science in general, not just Ariely’s work; but thanks to his books, Ariely’s work is widely known, so he serves as a paradigm case of what’s wrong with even the best of behavioral science today, and what we need to do about it.

Kindly bear in mind as you read this series: the critique is of the work, not the man.

Nasrudin was outraged. He ran back to the workshop teacher and said: “You told me the world was all peace and love, but that dog bit me! What’s that all about?”

The teacher shrugged. “Unfortunately, the dog didn’t take the workshop.”

Commentary: Vision – as in “visionary leadership” – has been a hot topic for a very long time. An ancient teaching says that “The world is as you see it.” This can be a very useful reminder, especially to organization leaders – after all, how can we change the way our organization relates to its customers, or engages its members in innovation, or creates a new product line, unless we first see it as a real possibility in our world? To change our world, perhaps we must first see the world as changed.

But Nasrudin may be reminding us that “the vision thing” only gets you half-way there. That ancient teaching applies all the way around: the world is as they see it, too. Unless your customers took the workshop with you, they will continue to expect the same relationship with your organization they have always had – unless you make a strong point of relating to them differently, in a way that creates value for them. Innovation and new product creation will stay as it has been unless you change the way your organization solicits and responds to ideas from your team and customers. People will respond to your lead as they always have unless you visibly change how you lead.

Vision – seeing the world as changed – is the crucial first step. But Nasrudin suggests you can’t then just wait for “them” to get the message. As a recent slogan puts it, you must be the change you see. After all, the dog didn’t take the workshop.

Contemplation: Where are you expecting the dog to have taken the workshop? How can you be the change, so they can change with you?

Self-report is notoriously unreliable; rigorous observation of actual behavior is very difficult to get right; creating comparable experimental and control groups takes more than a little ingenuity – and then you have the problem of replicable results.  Researchers who find methods that even partially steer around these known difficulties become heroes in their fields; their methods become standards, and genuine scientific knowledge can result.

But sometimes this ingenuity results in denying the obvious, and then we get silly science leading to bad theory. Some behavioral economics experiments, alas, fall into that category. They embody the Wizard of Oz fallacy: insisting that experimental subjects “Pay no attention to the man behind the curtain” when the curtain is standing open for all to see.

Case in point: Dan Ariely did a clever experiment on honesty involving refrigerators in MIT common rooms. Ariely is the Alfred P. Sloan Professor of Behavioral Economics at MIT, and a prolific researcher with colleagues from top universities around the world, who wrote an influential best-selling book Predictably Irrational: The Hidden Forces That Shape Our Decisions (2008). In this book he spells out, in a very readable form, some experiments he and his colleagues have done and the conclusions they reached as a result. There’s a lot of good stuff here, but unfortunately there’s also some breathtakingly bad stuff, and topping the list is the refrigerator experiment.

MIT common rooms, like most colleges, have refrigerators in which food and drink can be stored. The food is personal property, and strictly speaking taking food that is not yours is dishonest, but we all know that stuff left lying around tends to disappear over time.  Ariely sneaked a six-pack of Coke into a commons fridge and kept a diary of what happened.  As he puts it, “As you might expect, the half-life of Coke in a college dorm isn’t very long. All of them had vanished within 72 hours.”

But this was just to establish a baseline. Now came the real experiment, in which Ariely left in the refrigerator a plate of – are you ready for this? – six one dollar bills. And guess what – at the end of 72 hours, not one dollar bill had been taken!

This result may, or may not, surprise you but now look at Ariely’s explanation: “As we look at the world around us, much of the dishonesty we see involves cheating that is one step removed from cash.” He goes on to tee up his line of inquiry: “So what permits us to cheat when cheating involves non-monetary objects, and what restrains us when we are dealing with money? How does that irrational impulse work?”

Well, now. Before we go gallivanting off to pursue the roots of that “irrational impulse”, perhaps we should back up and look at that experiment again. On further review, the student’s “impulse” turns out to be far from “irrational”.

The essential scientific discipline, whether the outcome in question is a chemical reaction, the growth of a virus or the behavior of students in an MIT commons room, is careful control of variables. You do the same experiment twice, making very sure that the only difference between the two cases is the difference in the experimental variable; accordingly, any difference in outcome can be rationally attributed to the variable.

But unlike chemicals or viruses, students act in the light of their social circumstances. Change the circumstances, change the behavior – and the change may have nothing to do with your “experimental variable.”

Case in point: students using the common refrigerator are acting within a familiar set of social circumstances, which includes the presence of stray Cokes and the occasional liberation of one. It is within this context that their behavior makes sense. But now consider the moment when the student opens the refrigerator and sees a plate of one dollar bills. “What the …. ?” This is no longer the familiar social context. Dollar bills don’t belong in the refrigerator; something weird is going on and the student prudently does nothing.

Notice how rational this all seems. The behavior is fully explained by the change in social context; accordingly, we can conclude nothing about the supposed  “experimental variable”. Unless, of course, we insist that the students really didn’t notice anything odd about dollar bills in refrigerators, which is about as sensible as insisting that Dorothy pay no attention to the man behind the curtain as she looks right at him.

Shall we cut Ariely some slack here? Let’s assume he will chuckle and acknowledge all this, but he will say that the dollar bill experiment was meant to be a memorable stunt, pointing to the principle he then demonstrates  with sound research reported later in his chapter. OK?

Unfortunately, no. Ariely’s further research suffers from exactly the same Wizard of Oz problem (albeit not in such a glaringly obvious form). We’ll look at that in my next Post: “No, really, pay attention to the man behind the curtain!”

As reported in the January 14, 2010 edition of The Economist, Tanjim Hossain of the University of Toronto and John List of the University of Chicago “… worked with the managers of a Chinese electronics factory, who were interested in exploring ways to make their employee-bonus scheme more effective … At the beginning of the week, some groups of workers were told that they would receive a bonus of 80 yuan ($12) at the end of the week if they met a given production target. Other groups were told that they had “provisionally” been awarded the same bonus, also due at the end of the week, but that they would “lose” it if their productivity fell short of the same threshold.

“Objectively these are two ways of describing the same scheme.” But as it turned out, “The fear of loss was a better motivator than the prospect of gain (which worked too, but less well). And the difference persisted over time: the results were not simply a consequence of workers’ misunderstanding of the system.”Very interesting – and the results cry out for explanation, since classic economic theory would predict that the two schemes would produce the same results, and “positive organization” theorists would expect the reward to work better than the punishment. The researchers see this as essentially a “simple framing manipulation” using a notion from behavioral economics: “the value people attach to objects is affected by what they already have, and people abhor losses more than they like equivalent gains.” In other words, as The Economist opined: “Economists have always been advocates of using carrots and sticks. But they may not have emphasized appearances enough. Carrots, this research suggests, may work better if they can somehow be made to look like sticks.”

Well, no. I respectfully suggest that this is a great deal more than “appearances.” The fatal flaw of homo economicus stands out in high relief in that casual phrase “equivalent gains.” In Descriptive Psychology terms, this is classic Critic language. What is at issue here is understanding that behavior is engaged in by the Actor, not the Observer/Critic, and that what is “equivalent” to the Critic can be, and in this case is, categorically different to the Actor.

Here’s an alternative homo communitatus explanation: People act on what is real for them. They also act on what is possible for them, but “real” vs. “possible” is an important  distinction for an Actor. When a worker is given a bonus, even “provisionally”, that money becomes his – real money, a part of his real world – and he will treat it accordingly, including acting to prevent its loss. A bonus promised for performance at the end of the week is possible money.

To state explicitly the relevant homo communitatus principle: A person will value a real something over the “same” possible something.

Loss is not stronger than gain; real is more valuable than possible. And this is not a matter of “framing” or “appearances”; it is a matter of recognizing that Actor’s knowledge is not equivalent to Critic’s knowledge, which has been the fatal blind spot of classic economics all along.

(The crucial AOC concept and Actor’s knowledge are covered in depth in “At a Glance and Out of Nowhere” in the DPI papers, and, canonically, in Ossorio’s The Behavior of Persons. The full report of the research study can be found at: Tanjim Hossain & John A. List, 2009. “The Behavioralist Visits the Factory: Increasing Productivity Using Simple Framing Manipulations,” NBER Working Papers 15623, National Bureau of Economic Research, Inc.)

This is hardly a new or radical observation. A quick examination of the works of Ludwig Von Mises, Murray Rothbard, Herbert Simon and Gary Becker show that both the critique of homo economicus and substantive alternatives have been around in the works of respected economic theorists for over 50 years. Behavioral economists such as Amos Tversky and Daniel Kahneman have been expanding the model of the economic person since at least the 1970’s. And as New York Times columnist David Brooks points out, the economic events of 2008-2009 “exposed the shortcomings of the whole field. Economists and financiers spent decades building ever more sophisticated models to anticipate market behavior, yet these models did not predict the financial crisis as it approached. In fact, cutting-edge financial models contributed to it by getting behavior so wrong — helping to wipe out $50 trillion in global wealth and causing untold human suffering.” (March 25, 2010)

Why has is taken so long to abandon this plainly flawed approach? A 21^st century economist above all else must produce models and simulations, and good old homo economicus is so easy to use as a starting point. Those utility-maximizing algorithms are time-tested and give you genuine, hard numbers (even if, as it turns out, those hard numbers correspond to nothing at all in the real world we live in. I’ll have more to say about this in future posts.) As the gambler said as he returned to the craps table: “I know the game is rigged, but it’s the only game in town.”

There’s a new game in town. We can bury homo economicus in good conscience because we have its replacement: homo communitatus. Joe Jeffrey coined this term as the title of his 2010 paper, which uses the deep and highly developed conceptual resources of Descriptive Psychology to offer a starting point for economic persons. Homo communitatus (1) does justice to the  actual complexity of persons taking economic actions, and (2) provides a substantially more complex framework and toolkit for modeling and simulating economic matters.

No summation of the key points can do this paper justice (trust me, I’ve tried). It’s not obscure or hard to read; it’s just chock full of important conceptualization put together in an exciting and innovative way. You can find a pre-publication version of the  full paper at

http://www.descriptivepsychologyinstitute.org/HomoCommunitatis.pdf

But that’s just ante. Joe and I intend to articulate a new conceptual framework for economics. We’ve made a good start already. If you want to follow the development (and add your proverbial two cents) check out the DPI Center for Behavioral Economics at http://dpicbe.wordpress.com/