Sticky Luck

Seven:

The Psychology of Sticky Luck

Why do fund managers continue to thrive despite random, zero-mean performance?  Our brains and the human need for pattern recognition.  The pattern of heroic medicine.  The role of leadership and authority.  How demand and scale drive fees.  Bargains and bottom-fishing.  Famous patterns - heads and shoulders.  "Buy what you know."  Risk quality and the cost of living.

So far we have described how all fund managers’ and advisors’ performance could be nothing but luck, based upon the reasonable possibility of catching two or three very sticky stock-price series during a career.   My assertion could be disproved if we found a manager or advisor who consistently called both the ups and downs of the market with such frequency that it was beyond the reasonable possibilities of sticky luck.  Such calls would need to be unequivocal – a definitive buy/sell instruction clearly before the event, and generating superior profit evenly on both buys and sells with relentless consistency.  The balanced combination of both buys and sells is critical to weed out any closet rider of a sticky-luck series.  All calls would need to be counted – both the good and the bad – for example, if we are talking about a newsletter, it should list every call made, and apply no weighting to each unless it was applied at the time of the call.  I will not try here set down hard rules for a manager/advisor to pass such a test, but I think statisticians could easily agree that they would know a successful individual when they found the right data.  This book may be seen as a challenge for someone to accumulate such data that can be independently audited.

I would personally be pleased to find such data, if only to reassure myself that the financial industry that surrounds money management could provide the value for money expected of it, even if rarely achieved in practice.  Two reasons have led me to doubt that such data could exist.  The first is that any such capable manager or advisor would publish his or her list of calls by date, and have them independently audited for completeness and financial consequence.  Such a testament would surely have been published in some financial journal by now, offering conclusive proof that efficient-market theory is not valid.  My second reason is that I – not surprisingly from my earlier chapters – believe that efficient-market theory is intuitively compelling.  I think that those who do not find it compelling tend to underestimate – by a large degree of magnitude – the extraordinary amount of intellectual talent and technology that is invested in the market to arbitrage every money-making discrepancy as soon as it become apparent.

It is the thesis of this book that, to the extent that “gut feel” traders can squeeze quick profits from the market, they are taken very quickly and earned only by those who risk their capital in such enterprises.  Any manager or advisor who is paid by any fee arrangement which rewards them for luck, including sticky luck, has no incentive to invest in the brainpower needed to squeeze out momentary market inefficiencies.  Such fee arrangements obviously include percentages of assets managed, and subscription charges.  But, because of the nature of sticky luck, it also includes managers paid for beating market indices.  If I manage twenty funds on that basis, I can comfortably live on the ten that beat the averages, and either lose the remainder or manage them for free – knowing that some of them will perform better next time.  Moreover, if I am paid for performance, I have an incentive to increase my client’s risk exposure as much as possible, knowing that my personal upside is not balanced by my personal downside.  Risky luck will always pay me better on average, so long as it is my clients who are risking their money.

Confidence Trick?

To some readers, this thesis may amount to an accusation that fund management is little but a confidence trick, but I personally do not take that position.  Most respectable financial journals, which have no incentive to hide the truth about investment, take the quality of performance-seeking fund management seriously.  I am sure they are joined by many academics who also seek the truth.  They generally start from the perspective that no one can prove that fund managers cannot use their skills to consistently out-perform the market.   They also see the existence of such skills, as with many other demonstrable human capabilities which can appear extraordinary to the average person, as a matter of common sense.

Sticky luck is indeed difficult to absorb from a commonsensical perspective, as we are not used to seeing random results which form such chains of apparently meaningful sequences.  In the last chapter I attempted to provide a plausible explanation as to why this should occur.  Just as “predictable” market patterns can only be detected in retrospect, I attempted to explain this phenomenon by looking at causality in retrospect.   Starting with the fact of the market-changing, known result that has already driven the full progress of a stock or an index, we can look back on the gradual revelation of that result over a period of time, and the probabilities the market attached to its eventual success.  But, looking forward, we cannot know if future revelation will further confirm some possible or probable result that we currently estimate, or else reduce its likelihood.  This discontinuity between past and future, which keeps shifting as we move in time, runs counter to most natural experience.  Human common sense tells is that patterns must be traceable forward because they are causally connected to the past.

As intelligent animals, we have a built-in tendency to learn from patterns and rely upon the source of those patterns.  We are able to recognize identifying patterns in plants and animals so we can know them despite different sizes and shapes.  Evolution enhanced our brain’s pattern recognition so that we could comprehend the patterns of migrating herds, the seasons, rainfall, and plant growth.  Patterns in the environment were essential in our successful exploration and settlement of new lands.  More recently, our development of science is based upon the recognition of patterns, which are needed in order to create scientific hypotheses, and which in turn are tested in order to establish scientific theories from which most of today’s security, comfort and wealth are derived. Our advanced, imaginative pattern-recognition has made us the world’s dominant species, at least in terms of controlling our environment.

Nature’s Predictable Patterns

All these patterns come from Nature.  We have found Nature predictable in many ways, once we grasp its patterns.   When we know enough about natural patterns, we can predict elements of the future accurately, such as the rising and setting of the sun during the year.  Because of the complexity of interacting forces and agents, we cannot predict with such certainty the response of a body to a medical treatment, but we can determine a probable response.  We know the probable response by having observed patterns in other bodies’ responses, and by building a conceptual model of how the treatment affects parts of the body.   For example, we have found that administering a certain antibody provides statistical evidence that recipients are more resistant to a given disease, and we have a plausible model as to why this should be so.  Our pattern of resistance, combined with our model of what we think is going on, gives us confidence about the results of an inoculation program.  Maybe our model is flawed and has to be revised at some point, but we have some well-recognized and impartial ways of determining the treatment’s usefulness.

Now consider the former practice of bleeding a patient in order to cure his or her disease.  The practice continued for many centuries, and was still practiced by some physicians during the great influenza epidemic of 1918-19.   It is now well-accepted that bleeding has no beneficial effect whatever for most diseases, and is usually  harmful in its weakening of the patient, reducing the likelihood of recovery.  How could such a practice have continued for so long?   There was, of course, a hypothesis that seemed plausible at that time – since the role of blood in the body was not understood, it seemed that releasing its pressure would ease symptoms – it’s not difficult to imagine that many symptoms might be caused by this hot, thick liquid flowing in the body.  But what about the clinical evidence to support the hypothesis? 

It is not difficult to imagine many illnesses for which bleeding did not prove fatal and from which the patient rapidly recovers – many fevers would fall into this category, though they may look scary when at their worst.   A physician who practiced bleeding in a community might still enjoy an excellent recovery rate, and the loss of blood would seem to calm patients for a while by weakening them.  The practitioner would become well-known, and the unsuccessful physicians would hardly advertise their results for a broader picture of total clinical trials.  In fact, justice being rougher in those days, there would have been a tendency to select healthier, stronger patients anyway, and to hide failures whenever possible.  So it is easy to imagine celebrated physicians who owed their reputation to the practice of bloodletting.  The fact that one physician might practice for years on naturally recovering patients is an excellent example of sticky luck.  Such sticky luck would have been interpreted as a pattern, sufficient enough to be seen as endorsing the hypothesis, in the absence of anything else that could be found to work.  When we consider the medicines at the time, which included highly toxic ingredients such as mercury, the competition seems light.

The Common Sense of Heroic Medicine

We may view such practices as ignorant, but only in the sense that enough knowledge was unavailable to know better.  The practitioners were no less intelligent that physicians today – individual, raw intelligence has probably not progressed for thousands of years.  It is only our cumulative knowledge that has advanced, and we have all prior generations to thank for what we know today.  A plausible hypothesis, combined with sticky luck, would be sufficient to fuel the ambitions of some intelligent, persuasive men.  The tendency of others to follow the advice of such authorities is not unsurprising, especially in times of severe stress.  The sheer experience of the practitioner, compared with other folk, would trump any argument.  If the patient died, well, maybe the bleeding had not been started early enough or removed sufficient blood.   Even the death of someone as famous as George Washington through this practice had no effect upon its widespread use.

Would it be too outrageous to draw a parallel with fund managers and market advisors today?  We know that choices of different investments can make enormous differences to our wealth, whether we are a large pension fund or a person saving for retirement.  It is quite scary to know that simple choices, at the click of a mouse, can eventually double or halve our final nest egg.  To accept such choices as being random is probably as difficult as for the man or woman who sees their spouse in great pain and chooses to question to only remedy offered by a renown physician.

There is a vast amount of data, analysis and advice provided by banks, investment firms, insurance companies and all forms of media which must surely serve a valuable purpose for the investing public.  It is not difficult to be overawed by the weight of this sense of authority, even though there is not even superficially reassuring evidence that professional money managers out-perform the market, or that professional advisors have predictive power – even before we take into account the effects of sticky luck.  In the absence of good science we are, as with bloodletting, reassured by the trappings of authority and experience, born of constant practice, rather than evident success.

I also believe that, in such important areas of need, human beings actively seek out leadership in whatever form it can be found.  It is no doubt a feature of our ancient survival capabilities.  A caveman who had shown an exceptional ability to find food would no doubt be followed thereafter by his fellow cave-dwellers. Our ability to assess why that individual finds more food than others is secondary – if we waited for proof, we might die of hunger.   Probably it didn’t matter whether that person knew much more than the rest or was just the beneficiary of sticky luck through following the same tracks.  Perhaps survival depended as much on the group sticking together and following one set of orders.  For dogs and similar pack animals, this last explanation is known to be true.  It would not seem unreasonable for humans to follow similar instincts while rationalizing them into, “This person must have skills absent in the rest of us.”  Attaining such leadership may have been more a matter of appealing to such beliefs than of possessing the skills themselves.

As I have mentioned, questioning the value added by managers and advisors is not the same as accusing them of deceit.  If sticky luck can deceive the patient, it can also deceive the doctor.  The psychology of self-deceit is in fact probably much stronger in the case of managers or advisors who have spent their whole live studying all aspects of markets, foregone other fine careers that were available to them through their exceptional academic record and ambition, and who have been the beneficiaries of some impressive sticky luck.  When entering such a career they were no doubt unquestioning of the value of knowledge and insight to achieve long-term, superior returns.  Success of some type usually comes rapidly for such people to ever make it to the top.  There may have been severe losses when sticky luck went into reverse, but in their own minds this would have added to the gritty weight of their experience, helped paid their dues, and left battle-scars for a more interesting biography.  Though there may have been a short period when the question, “Am I just lucky?” haunted their work, the constant advice sought by colleagues and large, institutional clients would no doubt provide the answer, “They clearly don’t think so, and they’re bright enough to hold those jobs.”  Reinforcement of the credo is a natural process.

Then Add a Lot of Money

Especially when it is fueled by the vast sums paid for investment skills and advice.  Consider fund managers who charge one-half of one-percent of funds in order to follow an aggressive strategy which has an excellent track record against the S&P over ten years – a record that would be equaled by a significant number of players in the proverbial monkey-darts competition, but which is widely assumed to imply skill.  To investors, this half-percent (over, for example, the cost of a passive fund) would seem a small price to pay for skills that may have earned an additional 4-5% per year for the last few years.  Yet, on funds of $5 billion – a modest size – this discretionary income of $25 million a year could help pay for the most impressive research and presentation, and no doubt some original research which appeared to provide unique insights into current and long-term trends.  Now imagine funds under management of $50 billion or $500 billion, bearing in mind that today’s large companies exceed $50 billion each in market value.   The flow of income from such activity not only drives the most convincing show of patterns and conceptual “fundamentals,” it also reinforces to all involved that there must be real value here.  After all, it’s an open, competitive field.  Sadly, of course, so was bloodletting.  And competition based on sticky luck and slick advertising does not give investors value for money when they select active fund management.

I will again acknowledge here the obvious added value provided by traders who can respond quickly to new information and who, in effect, create market efficiency.  Without such individuals and a sufficient financial incentive, an efficient market could not exist.  But the financial incentive, in a competitive world, is just enough to reward these talented risk-takers. There would be no reason to share such rewards with the investing public, institutional or otherwise, if that public is content to pay for sticky luck.  The downside risk to market-makers is severe; those to fund managers relatively slight.  Also, as retail investors, we would have problems recognizing the difference between market-making and sticky luck, and being prepared to pay its fair value. 

I believe another psychological reason for retaining active fund management may be  one I touched on earlier – a misunderstanding of what market predictions actually mean.  Put simply, if the market is a kind of average of investors’ predictions, wouldn’t a good fund manager or advisor simply be someone who was better than average?  Put aside for the moment that all predictions, when weighted by their expected market returns, are random, so the concept of being “better” than average would have little useful meaning.   Put aside also that the market average is weighted by money invested, so should be dominated by the richest players who have greatest access to expertise.  Market predictions – which include any active management which is intended to allocate money to take advantage of future market conditions – have nothing to do with any supposed “true” value of investments.  They represent the predictor’s views on how the “market average” view will change over time.  Hence if a stock seems undervalued because of its P/E ratio, the predictor’s only useful input would be knowledge of the market’s future change of heart.  Not only does efficient-market theory tell us we can’t know this, but it seems strange that someone who disagrees with the market consensus today would predict that the market would move to share his or her opinion in the future.  Even if they could predict future unexpected information, such as surprise earnings, there is no reason why the market should share their response.

Bargains or Bottom-Fishing?

A common marketing tool used by advisors is to refer to “bargains.”  Bargains, as commonly understood, are items of worth sold at a discount.  If most people are selling a certain Rolex for $4,000 and I buy the same one for $2,500, I can reasonably call my purchase a bargain.  An analogy is made with the stock market when, after a 3-5% drop occurs, we commonly read of “bargain hunters” coming in to rescue the market.  This analogy is heavily misleading for at least two reasons:

·        Such a drop has, in recent times, typically followed annual rises of over 10% in prior years.  If a shop item had increased 30% in price over three years and then fallen 3-5%, would we refer to it as a bargain?

·        A large drop in a stock or index indicates the new price for all such items.  In our analogy, all the Rolexes would now be selling for $2,500, and comparable models have lined up with the same price differentials.  Suppose, instead of a Rolex, we are talking about a laptop, and the drop in price reflected older technology.  Would it still be a bargain?  Are last year’s models a bargain by virtue of being sold for less than this year’s models?  No, the market has determined the relative price.  And the negative news that drove down stocks 3-5% has, from the market’s perspective, just lowered their real value by that much.

We often come across some sound advice not to go “bottom fishing” in the market – in other words, buy all the out-of-favor stocks because their reduced prices have made them look cheap.  Under efficient-market theory, we know that the market-expected return has nor impoved.  However, quite often this advice is offered on the hypothesis that stocks that have fallen are in greater danger of falling further.  This gives us one of the many contradictions of received market wisdom – how do we square it with the professional activity of bargain-hunting, or the (now less favored expression, after the tech debacle) of “buying the dips”?   The advisor’s response is likely to be “it depends upon the fundamentals.”  But this argument becomes circular, as it retreats again into the concept of a “true value” that somehow the market is not understanding at this moment , but will within a limited time “get it” by bidding up those “bargain” stocks (or “smile” chart profiles, to use another appealing analogy) or bidding further down the trash of the bottom-fisher.

There seems to be an asymmetry is such advice, with bargains always outnumbering their opposite – the lack of an equivalent name for the opposite may be telling.  The lopsided buy-to-sell ratio of advisors is a symptom of this.  I think the saw-tooth nature of market patterns, where drops tend to be sharper than ascents, contributes to this tendency.   We see the more gradual ascent of stock value as somehow earned and its precipitous fall – though usually less than a full reversal and often quite minor – as being an “overreaction.”  Yet, as the tree example in the last chapter illustrated, a sudden break in a long sequence can confound expectations.  The saw-tooth is an intrinsic part of the sticky-luck phenomenon, not a bargain signal.

The received wisdom of market prediction can be slowly picked apart by lining up, in almost perfect balance, all the contradictory remarks that are offered as advice.  Their usual defense is vague “grain of truth” one, but this is of no practical help to the investor who must make decisions.  As mentioned in the first chapter, the investor is looking for actionable advice.  One major problem is that much advice, couched so conditionally as to be almost meaningless, is nevertheless interpreted by many eager investors as if it were a call to action.  (For example, “The market may be seriously undervalued,”  “We may have reached a temporary ceiling,” or “Don’t let fear keep you out of the market!”)  There is a high toleration of such silent nods and winks, as the conditionality is seen as being balanced advice, rather than as a “can’t-lose” position.  Ironically, many investors, frustrated with such opaqueness, choose instead to hand their funds over for discretionary management, assuming that they themselves must just not be getting the message.

Heads and Shoulders

One of my favorite psychology tricks, on both investors and advisors alike, is the famous “head and shoulders” pattern.  Like all chartists’ patterns, it is easy to find this one in great quantity, looking back on many stocks’ or indices’ histories.  The pattern involves an ascent to a temporary peak, followed by a plateau or temporary decline, followed by an ascent to an ultimate peak.  Thereafter it traces a similar pattern on the downside, ultimately collapsing to reflect, in whole or in substantial part, the original ascent.  Identification of head-and-shoulders is quite liberal, so the necessary requirements are limited to three peaks, the middle being the highest.  The purpose of identifying this pattern is, of course, to avoid the terrible journey down the far side.  We may not be able to identify this triple head until at least approaching the third peak but, if we get out then, we will have done well by accumulating and retaining most of the stock’s upside.

As with all such rules, the problem arises when we attempt to apply it consistently into the future.  Given the hi-tech collapse in 2002, and the preceding bubble, this period would seem to be very fertile for head-and-shoulders formations.  For many of the major hi-tech stocks, however, it is not so easy to find evidence of this phenomenon.  Those who believed the pattern to be a prime warming sign would have perhaps seen the 2001 NASDAQ peak as the first shoulder, and not been prepared for the almost continuous fall thereafter until the bottom.  Even if the 2001 peak was seen as the head, it would have been very difficult to have spotted any second shoulder.  We do find some examples, such as Yahoo!, though in a world of random patterns which included a big rise and big fall, it would be surprising if we did not.  Yet even here, it is not difficult for a skeptic to find more problems with the theory.  Look back at the rising years and find the temporary peak in 1998.  If we were at that point of time, with no knowledge of the future, it would be easy to view that peak as the head, with the shoulder just preceding it.  As the second peak fell away, there was a temporary pause.  The second shoulder?  To the pattern devotee, this would appear to have been perfect timing.  Yet the stock was poised at that moment to ascend much higher still. 

Of course, there are many heads and shoulders.  But there are just as many smaller brothers, with a head and shoulder but the other arm embracing the elder brother’s still higher shoulder.  Sticky luck ensures that we have any variation of any pattern type we care to conceive of, if we study enough historical charts.  What it won’t do is make it financially advantageous to assume that a certain type of pattern is taking shape.  Proponents have two defenses against this charge.  The first is to hedge their advice – “Be very careful.”  They didn’t tell you that the stock was headed south; they only warned that there is an “eerie similarity” to a pattern that occurred with other stocks. 

What could be wrong with such advice?  The fact that nothing could ever be wrong with such advice is an indication of its uselessness to an investor who must actually make decisions to buy, hold or sell.  If the advisor were saying anything useful, he or she would state that, given the pattern that had emerged to date, there would be a financial advantage in following a consistent form of action, even if it did not prove to be correct all the time.  In other words, there would be a financial advantage, as in poker, of playing the odds.  Such advice, though more relevant, cannot be supported without solid statistical evidence concerning all emerging, potential heads-and-shoulders and their ultimate outcomes.  Such research would need to be verified as comprehensive and unbiased by neutral parties.  It would certainly require an explicit definition of what is, and what is not, a qualifying pattern.  Has the reader seen such research claims for any market pattern?  Efficient-market theory says they does not exist, or would not pass scientific muster.

The other defense is the same as the bargain-hunting/bottom-fishing one.   It all depends on the circumstances, and how you apply the pattern analysis to them.  As with the conditional warning above, this simply moves the actionable advice back into the black box.  If the pattern itself is insufficient for us to make a decision, what specific additional information would trigger that decision?  If the advisor cannot define that trigger, then what is the value of the pattern itself?  Efficient-market theory tells us that there is no additional information that would make the pattern any less unhelpful than it already is, and the reluctance of an advisor to complete the predictive picture should confirm our suspicions. 

“Buy What you Know”

Another role played by advisors is to familiarize investors with a particular stock which they consider to be a “good buy.”  Even if this stock is not a red-hot tip, it may be positioned as an above-average investment opportunity, and the advisor’s role is not just to inform the investor of the pick but also to explain the nature of the investment.  Part of the philosophy here can be summarized as “buy what you know.”  It is often taken as sound advice, especially in the aftermath of the tech route, when obscure, niche-technology players were bid sky-high and then collapsed.  Within this class of advisors I include almost the entire popular media devoted to investment.

Later, we will discuss why knowing a company’s risk profile and sector role is important, but the idea that a retail investor can “know” a company enough to judge its superior or inferior investment potential is another example of illusory “market understanding.”  In a highly competitive world, there is no business which cannot suffer unexpected, severe setbacks by rivals, new technology, consumer shifts or economic upheavals.  The fact that we consume the products of a company tells us nothing about its potential for unexpectedly making or losing money.  As we have already discussed here, neither do any number of statistics about past performance, unless we believe we can beat the market to the punch within seconds.  I suspect the “buy what you know” dictum is another soft, entry sales-pitch of the advisory industry, creating a false sense of security in a volatile, random world.  When past sticky luck is used to illustrate the apparent wisdom of this investment approach, it joins market prediction as a fallacy.

I have questioned the value of the fund management and advisory industry, particularly in its attempt to deliver superior returns and provide simple solutions to market risk.  I have described efficient-market theory as implying the essential unpredictability of the market, and therefore the inability of an investor to expect a return in excess of any other investor, short of sheer news-reaction speed.  However, greater returns are of course possible through the acceptance of more risk.  Later in the book, I will also describe why accepting the “right” types of risk is a highly appropriate strategy for most investors.  This explanation may make efficient-market theory more palatable to many readers, who may have concluded so far that it implied all investments are of equal value to everybody.  Nothing could be further from the truth. 

As I have mentioned earlier, perhaps the biggest psychological barrier to understanding the stock market is the understanding of risk.  There are two sources of confusion: one is the potential size of risk, and the other is the quality of risk.  My description of the fractal nature of market movements suggests that the market carries potentially huge risk even in the short term, though the chances of a sustained collapse may be small.  It is not inevitable that the market will recover from a collapse.   I believe too many people investing in the market do not account for these possibilities, as they imagine some ultimate self-adjusting mechanism, simply because markets have always eventually recovered in the past.  I believe the market itself fully reflects this risk in its valuations, which is one of the reasons that bond holders are willing to accept historically lower returns.

Risk Quality

Quality of risk is much more complex to describe, and is covered in detail in the final chapter of this book.  First, I need to dig more deeply into the subject of market-expected return, developing a model which can also help explain many phenomena of the market.  The next chapter covers this subject.  But I should spend a little more time here on risk quality, if only because I will continue to use the term “risk-free rate of return.”  This concept is the return you can be sure of earning in dollar terms over a given period of time.  (Also, for now, I will not define that period of time, just assuming it to be the same length for any set of investments I am comparing.)  However, risk, or lack of it, should typically not be measured in dollar terms for most investors.  For individuals saving for retirement, for example, the risk is not having sufficient funds to enjoy a happy retirement.  In this context, the cost of living is a vital factor, rather than a fixed number of dollars. 

My discussion on the market-expected rate of return assumes that all investors are attempting to maximize their personal expected rate of return.  No matter how irrational we may think investors behave, I think the attempt to maximize is a reasonable premise for analysis.  Though we are ignoring limitations such as the requirements for investments to be socially responsible, their impact is probably very minor.  If we are comparing dollar-denominated investments, it is therefore useful to compare their dollar-denominated returns.  If my argument proves convincing, the reader will conclude that all dollar-denominated returns are equivalent on a market-expected basis.  In the absence of evidence that manager and advisors can improve upon that market-expected return, there is no point in paying anyone to try.  Instead, investors must take a personal perspective on risk, and choose portfolios that cater best to their preferred risk profiles. Once chosen, this portfolio makes the market-expect returned largely irrelevant.

Why irrelevant?  Because it is no more relevant that the expected score of 3.5 when a gambler throws a single die once.  He or she will not get 3.5 – he will get either more or less.  Similarly, we live only one life, which will contain one set of events.  The weighted probabilities of countless possible events will not determine our actual return – only actual events will do so.  Instead, we plan for what we consider is probable – a significantly higher return than the risk-free one, if we so choose – and prepare for the worst, or at least some of the less pleasant possibilities.  That is, unless we don’t mind facing the small chance of disaster.  But in that case, it would make no sense to buy any insurance that the law doesn’t force us to buy.  There is no excuse for inconsistency when it comes to risk.