Mar

25

To expiate for one of my greatest sins, the tendency to refrain from holding a position with a profit, I have been reading many articles on Loss Aversion, in marketing, law, and psychology. Here is a simple one [18 page pdf]  that covers many of the basics. 

How can one's decisions be less harmful and more rational? When you're given a good, you often will sell it at at least twice the price that you would buy it for. This explains so many things like bid asked spreads and free trial offers. I would like your input on how to reduce the harmful effects of loss aversion. How does it relate to risk aversion for example?

A commenter writes:

If managing money, will it work to objectively identify clients with those biases? (Larger sample size to train on.) Then one day see yourself as a client.

David Lililenfeld writes: 

Changing behavior is one of the hardest things one can do, but as most successful marketers will tell you, it can be done in almost any circumstance. There are apps for the iPhone (I can't speak for Android) which have succeeded in getting people to exercise or lose weight. Perhaps you might adapt one of them to suit your need.

a commenter writes:

Creating a graphic representation of the effects of loss aversion manifested by repeated failure to properly execute a specific trading system, especially holding onto the inevitable losers for the duration while grabbing profits on positive MTM positions before they fully develop, should drive the point home.

Viewing the terrible transformation of the equity curve should be enough to cure the speculator of the detrimental affliction.

Richard Owen adds: 

The simulations of St. Petersburg's bet are explained according to a computation of concave utility or an asymmetric loss reaction plus max payout. It seems that rational expectations man creeps into even psychological bias literature and perhaps indicates the sometimes impractical nature of academic analysis. Why is the most likely explanation not someone being offered the bet, thinking "I don't know what you're talking about, but I'll wager anyhow for fun", and thus their average wager was some median of $5, $10, $20, $50, people's intuitive level for mouth bets.

Victor Niederhoffer adds: 

Yes. If loss aversion is pervasive, then it should show up in regularities relating to price moves. The situation is complicated in futures where one person's long profit when price goes up is the short's loss. The endowment effect which is caused by loss aversion or the tendency to connect with what you own, could lead to holding something too long. The reference point effect, which is that people base their decisions on where they are, a variant of holding onto the status quo is also a factor. When there is a profit, a different type of endowment effect plays then when there is a loss. Especially when there has been a big loss and it turns into a profit, the loss aversion effect is greatest I believe. I have been trying to quantify various regularities relating to these sometimes conflicting and ever changing effects in the market, and perhaps it will not be a total loss. Thanks for the suggestions that my confession has elicited, and I apologize to all of you for my transgressions in this regard in the past. Had I not made such transgressions so often in the past I believe I would be a wealthy man. I am also open to the stages of confession of a sin from alcoholics anonymous, and perhaps I am at one of the stages by confessing my sins to a support group hopefully of ones who have conquered this terrible tendency.

Jeff Watson adds: 

When I get in a trade, I look at the size of what I have on, what the market looks like, is the market going to have the wind at my back, and then I pre-determine where I will take a profit, add more position, or at what specific price will I take a quick hit and get on with it. I do not deviate from that loss plan, ever. I might get getting nickeled, dimed and vigged to death, but the slow torture the market gives me is much preferable to trading by the seat of my pants, being consistent leaker, or blowing out spectacularly like I've seen so many do.. Taking my first loss quickly is the only rule in the book I don't break. It allows me to sleep well, and sleeping is important to me. That being said, whenever I'm in any marker, I always pay too much for it or sell it too cheaply. It's worse whenever I make a foray outside the grains, since other markets are outside my comfort zone, trading in them for me is strictly gambling, and I'm not very good at gambling.

Ralph Vince writes: 

The St Petersburg Paradox is a false argument, one that assumes that rational human beings arrive at conclusions based on mathematically "correct" asymptotic answers. But we do not operate this way, rather, in an amazing an beautiful twist, our minds operate based on a notion of horizon, a finite period in which to operate.

This is the reason you will board an airplane to get to your destination (bet case) even though you know it could crash (worst case). The mathematical expectation for such (the same formulated expectation used in assessing the St Petersburg Paradox — the sum of the outcomes times their probability) would assign an outcome so infinitely negative to the prospect of death, that even though the probability were so minute, the expectation would still be so negative that you would never board that plane.

Yet, in order for that minute expectation to manifest would require X trials, X flights, and X is a VERY large number relative to how many flights you will take in your lifetime — even for a commercial pilot.

And we innately know this! This is the very decision-making mechanism that allowed our early ancestors to descend from the high branches and onto the savanna at the outset of our long journey. Asymptotically, it was a bad bet. To innate human reasoning, it was a god bet.

So especially if we are talking about someone "betting for fun" or participating for the hell of it, we can not look at things through a lens focused on the infinite asymptote, but must focus on a finite horizon. The St Petersburg Paradox itself now solves readily.

Christopher Tucker writes: 

For my part, I find that on the frequent occasions that I kill a position to early, it is because I am paying attention to the wrong data. I am either paying too much attention to either:

- the profit or loss - focusing on the number of dollars or the percentage of capital
- the wrong time frame

The time frame is a huge one for me - I enter trades using criterion in a longer frame and then drill down to look for an opportune moment in a shorter time frame to pull the trigger. The problem for me is that I tend to remain looking at the shorter time frame and exit too soon or with a stop that is too tight. It is very much akin to missing the forest for the trees.

Much the same can be said in my main business of air traffic control. Rookies tend to pay attention strictly to proximity, which is critical, but tend to not pay attention to speeds, vertical speeds or properly anticipated speeds based on a solid knowledge of aircraft types or the current terrain. A Boeing 747-800 may fly at a reduced speed while climbing to get above an area of turbulence, but the moment he is above it or levels off, you can bet heavily that he will accelerate and very quickly at that. By the same token, high performance business jets like Gulfstreams and Learjets can be counted upon to climb rapidly (sometimes extremely so), especially if they have been in level flight for a while and are itching to get higher.

The key for me is paying attention to the right things. Profit and loss are important, time frames are important, but they may not be the best indicators of the overriding theme.

Steve Ellison writes: 

We have many times discussed path dependency. As the paper says, "[R]eceipts are naturally encoded, and responded to, as gains and losses from a reference point. So a pauper who wins $10 million in a lottery may be ecstatic with her staggering $10 million in net worth, while a dot-com tycoon whose $100 million paper fortune shrinks to only $10 million in a week may kill himself, distraught over having only a pathetic $10 million in net worth."

As an example from a different field, consider a baseball pennant race. If team B was 7 games behind team A in August, but has pulled into a tie by late September, what are the teams' psychological states? Team A's clubhouse has the atmosphere of a funeral, while team B is brimming with excitement and confidence. I have not put pencil to paper, but I grew up a fan of the Boston Red Sox, who were usually team A, and it seemed that team B nearly always prevailed in such situations. Logically, the recent past should count for nothing because the teams have exactly the same records, and only the future games will determine the winner.

Gary Rogan writes:

Why isn't this a choice between two diametrically opposed systems: either a buy and hold, where there is no special characteristics attached to the winner or the losers, or a mathematical system that has some predetermined dispositions at any time for a winner and a loser based on some quantitative criteria. When it's something in between, that ambiguous choice and regret enter the picture.

Dan Grossman writes:

Probably not many here, but there are those of us who at times own securities until they are long-term, and sometimes refrain from taking profits early because of the pressure to wait for long-term gain tax treatment.

Perhaps a retrospective study of such an investor's stock purchases that are significantly up after (say) three months, comparing annualized return of those stocks sold after one year, compared to those sold short-term.

Sushil Kedia comments:

Loss Aversion emanates from the desire to enjoy profits without having to respond to losses. It is a manifestation of an ill that returns can be generated without undergoing adverse excursions. Loss Aversion is placed within a fantasy that the majority suffers much more frequently.

Risk Aversion respects the law of diminishing marginal utility of risk. It acknowledges well that returns are feasible only upon a readiness to respond to adverse excursions.

The unwillingness to respond to adverse excursions results in all of the sins that traders can eventually end up doing, including losing more than what they are entitled to on any one adventure or set of adventures in the markets. Killing a position too soon is also a manifestation of loss aversion only, with a variant that one is averse to losing the profits away. I would surmise the opposite of this sinful loss aversion is not loss seeking behaviour, but merely loss acknowledging behaviour.

Risk aversion seeks to undertake an additional unit of risk only if a larger unit of return can be added. A relationship similar in nature to diminishing marginal utility thus makes for a maxima that is likely to be the optimal level of risk for a method, strategy, approach, history, future or portfolio. Loss aversion however is not only playing against the odds, but is rather an exponentially growing risk relationship. For an example if a 20% loss needs to be evened without being closed a 25% favourable move is required and so on and so forth [x loss needs x/(1-x) gain immediately to get even]. There is no optimal point possible in this relationship and the approach can and does take the indulgent down to zero.

For now, I would say Loss Aversion is driven by ego as to what the market should do, whereas Risk Aversion is driven by humility as to what the market may do.

a commenter writes: 

Continuing to hold a profitable position (long or short) is trend following: the expectation of movement in the same (profitable) direction.

To the extent there is aversion to losing profits (and closing out too soon), this risk should be compensated and may in part explain profitable trends.

Craig Mee writes: 

I think it's hard to naturally run trades when mean reversion is so frequent and we can't guarantee when the trend will change and the law of ever-changing cycles kicks in. So to counter that, it makes sense to trade in multiplies. Take half risk off almost straight away on the first push, bringing the chance of a scratch trade at worst, rather than a loss higher on the curve… though if the market does run, particularly in fx or higher range markets, subsequent setups can present opportunities for adding further. So a case of give and take I suppose, when from the outset nothing is guaranteed.

Phil McDonnell writes: 

We have had numerous discussions on this venue regarding stop losses. Part of the surprise from those discussions is that using a stop loss will double your odds of having a loss in the amount of the stop loss.

However the same is true for a profit target. Using a profit target will double your probability of having a gain equal to the target gain. The reason for both phenomena is that in a random walk half of all such trades will get reversed after hitting the target or the stop. The fancy name for this is the Reflection Principle.

Jeff Rollert responds to Jeff Watson: 

I have to tell you, having the wind at your back is not the fastest point of sail on a boat, nor the least risky if the waves are coming from your stern. Please google and study canoe sterns for a discussion. I bring this up only in context of its use.

I find many sailing "descriptions" along with those in climbing mountains to be misapplied. In both cases the getting into the trade is over weighted vs the exit.

As Hillary implied, he point of climbing is to live after coming down the mountain. That math would overweight the sell decisions.

I tip my hat to Lowe passing recently. Like Hillary and Norgay, he had balls and lived to tell the story. I highly recommend "Touching My Fathers Soul" in that spirit…who unlike Mallory was ill prepared and long of ego.

Gambling in climbing, is being unable to leave the mountain without summiting. Sailing has its parallels. So do the Marines.


Comments

Name

Email

Website

Speak your mind

2 Comments so far

  1. freerider1 on March 26, 2013 6:59 pm

    The problem of winning:
    While more you win, more dependent you become on winning as it becomes the “normal” outcome expected. Hence, when you take a loss it tends to hurt quite a lot. In addition to that, I guess, your trading confidence goes from normal to overconfident, hence more probability for a reversal. And, since winning is tempting, there is also an additional risk: you might as well scale up the bets and get exposed too much to kurtosis risk.

    The solution:
    More losing, hence expose yourself positively to winning.

  2. Ramakrishna Shanker on March 26, 2013 11:07 pm

    St Petersburg can be solved without resorting to any “utility” functions or “irrational” behavior. It has to do with ergodicity. St Pete is highly non-ergodic. ie. The ensemble average is not equal to the time average. A simple calculation of the time average ( which is what most of us care about) gives the correct answer. In other words, if you play the game repeatedly and want to have a positive return, what is the most you would pay to play?

Archives

Resources & Links

Search